Linear Programming Computation 2nd Edition Ping-

Qi Pan download

https://ebookmeta.com/product/linear-programming-computation-2nd-
edition-ping-qi-pan/

Download more ebook from https://ebookmeta.com

We believe these products will be a great fit for you. Click
the link to download now, or visit ebookmeta.com
to discover even more!

Advances in Optimization and Linear Programming 1st

Edition Ivan Stanimirovi■

https://ebookmeta.com/product/advances-in-optimization-and-
linear-programming-1st-edition-ivan-stanimirovic/

Linear Programming Foundations and Extensions 5th

Edition Robert J. Vanderbei

https://ebookmeta.com/product/linear-programming-foundations-and-
extensions-5th-edition-robert-j-vanderbei/

Linear Programming for Project Management Professionals

1st Edition Partha Majumdar

https://ebookmeta.com/product/linear-programming-for-project-
management-professionals-1st-edition-partha-majumdar/

Bioinformatics and Computational Biology 1st Edition

Hamid R. Arabnia

https://ebookmeta.com/product/bioinformatics-and-computational-
biology-1st-edition-hamid-r-arabnia/
Advances in Postharvest Management of Cereals and
Grains 1st Edition Dirk E. Maier (Editor)

https://ebookmeta.com/product/advances-in-postharvest-management-
of-cereals-and-grains-1st-edition-dirk-e-maier-editor/

Abdication of the Sovereign Self The Psycholinguistics

of Invalid Synthetic Propositions 1st Edition Andrew
Spano

https://ebookmeta.com/product/abdication-of-the-sovereign-self-
the-psycholinguistics-of-invalid-synthetic-propositions-1st-
edition-andrew-spano/

Hybrid Forgotten (Hybrid Shifters Book 10) 1st Edition

Romy Lockhart

https://ebookmeta.com/product/hybrid-forgotten-hybrid-shifters-
book-10-1st-edition-romy-lockhart-2/

Lock the Door Jane Holland

https://ebookmeta.com/product/lock-the-door-jane-holland/

Antimicrobial Photodynamic Therapy 1st Edition Busi

Siddhardha Ram Prasad

https://ebookmeta.com/product/antimicrobial-photodynamic-
therapy-1st-edition-busi-siddhardha-ram-prasad/
Introduction to Positive Media Psychology 1st Edition
Arthur A Raney Sophie H Janicke Bowles Mary Beth Oliver
Katherine R Dale

https://ebookmeta.com/product/introduction-to-positive-media-
psychology-1st-edition-arthur-a-raney-sophie-h-janicke-bowles-
mary-beth-oliver-katherine-r-dale/
Ping-Qi PAN

Linear
Programming
Computation
Second Edition
Linear Programming Computation
Ping-Qi PAN

Linear Programming
Computation

Second Edition
Ping-Qi PAN
Department of Mathematics
Southeast University
Nanjing, China

ISBN 978-981-19-0146-1 ISBN 978-981-19-0147-8 (eBook)

https://doi.org/10.1007/978-981-19-0147-8

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore
Pte Ltd. 2023
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Dedicated to my parents, Chao Pan and
Yinyun Yan; my wife, Minghua Jiang; my son,
Yunpeng; my granddaughter, Heidi
Preface to First Edition

Linear programming (LP) founded by Dantzig (1948–1951b) might be one of the

most well-known and widely used mathematical tools in the world. As a branch
of optimization, it serves as the most important cornerstone of operations research,
decision science, and management science.
This branch emerged when the American mathematician George B. Dantzig
created the LP model and the simplex method in 1947. The computer, emerging
around the same period, propelled the development of LP and the simplex method
toward practical applications. As a basic branch, LP orchestrated the birth of
a number of new branches, such as nonlinear programming, network flow and
combinatorial optimization, stochastic programming, integer programming, and
complementary theory, etc., and activated the whole field of operations research.
A prominent feature of LP is given by its broad applications. Closely related
to LP, many individuals have made pioneering contributions in their respective
fields. In the field of economics, in particular, Russian-American economist Wassily
Leontief took the 1973 Nobel Economic Prize for his epoch-making contribution
to quantitative analysis of economic activities. The academician L. V. Kantorovich
of the former Soviet Academy of Science and American economist Professor T. C.
Koopmans won the 1975 Nobel Prize for their optimal allocation theory of resources
using LP. The same prize was also given to Professors K. Arrow, P. Samuelson, H.
Simon, and L. Herwricz, several decades later when they paid close attention to LP
at the starting days of their professional careers.
The simplex method has also achieved great success in practice. As it is known,
its applications to many fields, such as economy, commerce, production, science and
technology, and defense and military affairs, etc, have brought about astonishing
economic and social benefits. It is recognized as one of The Ten Algorithms in the
Twenty Century (IEEE2002; see Cipra (2000).
After more than 70 years since its birth, LP is now a relatively mature but still
developing branch. Nevertheless, there exist great challenges. The importance of
large-scale sparse LP models is nowadays enhanced further by globalization. The
everyday practice calls upon the research community to provide more powerful
solution tools just to keep up with the ever-increasing problem sizes. Therefore, this

vii
viii Preface to First Edition

book does not only present fundamental materials but also attempts to reflect the
state of the art of LP. It has been my long-lasting belief that research, in operations
research/management science, in particular, should be of practical value, at least
potentially. The author, therefore, focuses on theories, methods, and implementation
techniques that are closely related to LP computation.
This book consists of two parts. Part I mainly covers fundamental and conven-
tional materials, such as the geometry of the feasible region, the simplex method,
duality principle, and dual simplex method, implementation of the simplex method,
sensitivity analysis, parametric LP, variants of the simplex method, decomposition
method, and interior-point method. In addition, integer linear programming (ILP),
differing from LP in nature, is also discussed not only because ILP models can be
handled by solving a sequence of LP models but also because they are so rich in
practice and form a major application area of LP computations.
Part II presents published or unpublished new results achieved by the author
himself, such as pivot rule, dual pivot rule, simplex phase-1 method, dual simplex
phase-I method, reduced simplex method, generalized reduced simplex method,
deficient-basis method, dual deficient-basis method, face method, dual face method,
and pivotal interior-point method. The last chapter contains miscellaneous topics,
such as some special forms of the LP problem, approaches to intercepting for primal
and dual optimal sets, practical pricing schemes, relaxation principle, local duality,
decomposition principle, and ILP method based on the generalized reduced simplex
framework.
To make the material easier to follow and understand, the algorithms in this book
were formulated and illustrative examples were worked out wherever possible. If
the book is used as a textbook for upper-level undergraduate/graduate course, (Part
I) may be suitable to be as basic course material.

Acknowledgments

At this very moment, I deeply cherish the memory of Professor Xuchu He, my
former mentor at Nanjing University, aroused my interest in optimization. I honor
the Father of linear programming and the simplex method, Professor George. B.
Dantzig for the encouragement given during the 16th International Symposium on
Mathematical Programming, Lausanne EPFL., in August 1997. I am very grateful
to Professor Michael Saunders of Stanford University for his thoughtful comments
and suggestions given in the past. He selflessly offered the MINOS 5.51 package
(the latest version of MINOS at that time) and related materials, from which my
work and this book benefited greatly. MINOS has been the benchmark and platform
of our computational experiments. I thank Professor R. Tyrrell Rockafellar of the
University of Washington, Professor Thomas F. Coleman of Cornell University, and
Professor Weide Zheng of Nanjing University for their support and assistance. I
would also like to thank the following colleagues for their support and assistance:
Professor James V. Burke of the University of Washington, Professor Liqun
Preface to First Edition ix

Qi of Hong Kong Polytechnic University, Professor Lizhi Liao of Hong Kong

Baptist University, Professors Leena Suhl and Dr. Achim Koberstein of Paderborn
University, Professor Uwe Suhl of Freie University, and Dr. Pablo Guerrero-Garcia
of the University of Malaga.
This book was supported partially by projects 10871043 and 70971136 of the
National Natural Science Foundation of China.

Nanjing, China Ping-Qi Pan

December 2012
Preface to Second Edition

Since its birth, linear programming (LP) has achieved great success in many fields,
such as economy, commerce, production, science, and technology, and brought
about amazing economic and social benefits. After 70 years of development, LP
now becomes a relatively mature branch in OR/MS. Nevertheless, the academic
community faces a major challenge of the growing demand, which needs more
powerful and robust tools to deal with large-scale and difficult problems.
To meet the challenge, this book draws materials from a practical point of
view, focusing on theories, methods, and implementation techniques that are
closely related to LP computation. Its first edition consists of two parts. Roughly
speaking, Part I covers fundamental materials of LP, and Part II includes pub-
lished/unpublished new results achieved by the author.
Little attention was received via the author’s ResearchGate Web page until Shi,
Zhang, and Zhu published their book review in the European Journal of Operational
Research (June 2018). Since then, the book quickly attracted considerable attention
from academic communities around the world. As of November 6, 2022, the top ten
chapter reads are as follows:

Chapter Reads
1. Duality Principle and Dual Simplex Method (Part I) 15419
2. Simplex Feasible-Pointx Method (Part II) 7861
3. Integer Linear Programming (ILP) (Part I) 7340
4. Implementation of the Simplex Method (Part I) 3284
5. Dual Simplex Phase-I Method (Part II) 2613
6. Simplex Method (Part I) 2576
7. Variants of the Simplex Method (Part I) 2328
8. Geometry of the Feasible Region (Part I) 1951
9. Pivot Rule (Part II) 1844
10. Simplex Phase-I Method (Part I) 1506

xi
xii Preface to Second Edition

The preceding data are consistent with corresponding downloads from the
Springer website.
Surprisingly enough, the chapter ”Dual Principle and Dual Simplex Method” not
only received the maximum number of reads as a whole but also almost weekly,
even if it is devoted to such a classical topic. This might be because the discussions
are comprehensive, compared with popular literature.
Proposed for the first time, the Feasible-Point Simplex Method, which appeared
as the final section of the chapter ”Pivotal Interior-Point Method,” received the
second-highest reads as a whole. It achieved this in a short period of 10 weeks.
Indeed, supported by the solid computational results, the method itself might exceed
all expectations.
The final chapter of Part I, ”Integer Linear Programming (ILP),” usually received
the second-highest reads weekly. This happened, I guess, because many researchers
are interested in the newly proposed controlled-branch method and controlled-
cutting method, which have potential applications for developing new ILP solvers.
The author’s main contributions seem to have not been fully valued by now,
including most pivot rules, reduced and D-reduced simplex methods, deficient-basis
and dual deficient-basis methods, face and dual face methods, and the decomposi-
tion principle, included in Part II. This is somewhat surprising since these methods
are supported by solid computational results, except for the newly introduced
reduced and D-reduced simplex methods, and the decomposition principle. In the
author’s view, the reduced simplex method is particularly noteworthy. It is the
first simplex method that searches along the objective-edge. In contrast to existing
simplex methods, it first determines pivot row and then pivot column, and does
so without any selection essentially. As a dual version of it, the D-reduced simplex
method shares similar attractive features, through determining pivot column first and
pivot row later. As for the decomposition principle, it allows for solving arbitrarily
large-scale (even dense) LP problems, in essence, giving a glimmer of light on some
other types of separable large-scale problems. Time will likely clarify the value of
these methods vs other LP solvers.
We are also optimistic about the face and dual face methods. Without exploiting
sparsity, the original methods were implemented using orthogonal transformation
with favorable computational results reported. In the second edition, we added
two chapters devoted to new methods with LU factorization for sparse computing.
Indeed, this is a very natural development which the author cannot help himself
from breaking into.
Some changes on chapters and sections were made with corrections and improve-
ments, and the whole second edition was organized into two parts. Roughly speak-
ing, Part I (Foundations) contains Part I of the first edition, and Part II (Advances)
includes Part II. The Simplex Feasible-Point Algorithm was improved, and removed
from the chapter ”Pivotal Interior-Point Method” to form an independent chapter
with its new title “Simplex Interior-Point Method,” since it actually represents a
new class of interior-point algorithms, which can be transformed from the traditional
simplex algorithms. The title of the original chapter was changed to “Facial Interior-
Point Method,” since the remaining algorithms represent another new class of
Preface to Second Edition xiii

interior-point algorithms, which can be transformed from the normal interior-point

algorithms. In particular, the chapter “Integer LP” was rewritten in great gains of
the introduction of the objective cutting. Another exciting improvement was the
reduced simplex method. The original derivation of its prototype was presented in
a chapter with the same title, and then transformed into the so-called “improved”
one in another chapter. Fortunately, inspired by the bisection simplex method, we
recently found a quite concise new derivation, so we can introduce it in a single
chapter now.
Finally, I am grateful to many individuals for their valuable comments and
suggestions on the first edition of this book. In particular, I appreciate comments
and encouragement given by Dr. Y.-Y Shi, Professor L.-H Zhang, and Professor
W.-X Zhu.

Nanjing, China Ping-Qi Pan

November 2022

References

1. Cipra BA (2000) The best of the 20th century: editors name top 10 algorithms.
SIAM News 33:1–2
2. Murtagh BA, Saunders MA (1998) MINOS 5.5 user’s guid. Technical report SOL
83-20R, Department of Engineering Economics Systems & Operations Research,
Stanford University, Stanford
3. Shi Y-y, Zhang L-H, Zhu W-X (2018) A review of linear programming compu-
tation by Ping-Qi Pan. European Journal of Operational Research 267(3):1182–
1183
Contents

Part I Foundations
1 Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3
1.1 Error of Floating-Point Arithmetic . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4
1.2 From Real-Life Issue to LP Model . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6
1.3 Illustrative Applications . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 9
1.4 Standard LP Problem.. . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 14
1.5 Basis and Feasible Basic Solution .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 17
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 22
2 Geometry of Feasible Region .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 23
2.1 Feasible Region as Polyhedral Convex Set . . . . .. . . . . . . . . . . . . . . . . . . . 24
2.2 Interior Point and Relative Interior Point . . . . . . .. . . . . . . . . . . . . . . . . . . . 28
2.3 Face, Vertex, and Extreme Direction . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
2.4 Representation of Feasible Region . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 42
2.5 Optimal Face and Optimal Vertex . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 44
2.6 Graphic Approach .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 46
2.7 Heuristic Characteristic of Optimal Solution . . .. . . . . . . . . . . . . . . . . . . . 47
2.8 Feasible Direction and Active Constraint . . . . . . .. . . . . . . . . . . . . . . . . . . . 50
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 54
3 Simplex Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 57
3.1 Simplex Algorithm: Tableau Form . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 57
3.2 Getting Started.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 68
3.3 Simplex Algorithm .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 74
3.4 Degeneracy and Cycling . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 80
3.5 Finite Pivot Rule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 84
3.6 Notes on Simplex Method . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 91
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 93
4 Implementation of Simplex Method .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 95
4.1 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 95
4.2 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 97

xv
xvi Contents

4.3 LU Factorization of Basis . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 99

4.4 Sparse LU Factorization of Basis . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 102
4.5 Updating LU Factors .. . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 106
4.6 Crash Procedure for Initial Basis . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 109
4.7 Harris Rule and Tolerance Expending.. . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 111
4.8 Pricing for Reduced Cost . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 113
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 117
5 Duality Principle and Dual Simplex Method . . . . . . . .. . . . . . . . . . . . . . . . . . . . 119
5.1 Dual LP Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 119
5.2 Duality Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 122
5.3 Optimality Condition.. . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 125
5.4 Dual Simplex Algorithm: Tableau Form . . . . . . . .. . . . . . . . . . . . . . . . . . . . 129
5.5 Dual Simplex Algorithm .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 132
5.6 Economic Interpretation of Duality: Shadow Price .. . . . . . . . . . . . . . . . 136
5.7 Dual Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 137
5.8 Bilevel LP: Intercepting Optimal Set . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 142
5.9 Notes on Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 146
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 147
6 Primal-Dual Simplex Method . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 149
6.1 Mixed Two-Phase Simplex Algorithm . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 149
6.2 Primal-Dual Simplex Algorithm.. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 152
6.3 Self-Dual Parametric Simplex Algorithm .. . . . . .. . . . . . . . . . . . . . . . . . . . 158
6.4 Criss-Cross Algorithm Using Most-Obtuse-Angle Rule . . . . . . . . . . . 162
6.5 Perturbation Primal-Dual Simplex Algorithm . .. . . . . . . . . . . . . . . . . . . . 166
6.6 Notes on Criss-Cross Simplex Algorithm .. . . . . .. . . . . . . . . . . . . . . . . . . . 170
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 170
7 Sensitivity Analysis and Parametric LP. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 173
7.1 Change in Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 173
7.2 Change in Right-Hand Side .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 176
7.3 Change in Coefficient Matrix . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 178
7.3.1 Dropping Variable . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 178
7.3.2 Adding Variable . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 180
7.3.3 Dropping Constraint .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 182
7.3.4 Adding Constraint . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 184
7.3.5 Replacing Row/Column .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 189
7.4 Parameterizing Objective Function .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 192
7.5 Parameterizing Right-Hand Side . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 196
8 Generalized Simplex Method. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 199
8.1 Generalized Simplex Algorithm .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 200
8.1.1 Generalized Phase-I . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 207
8.2 Generalized Dual Simplex Algorithm: Tableau Form . . . . . . . . . . . . . 207
8.2.1 Generalized Dual Phase-I . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 212
Contents xvii

8.3 Generalized Dual Simplex Algorithm .. . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 213

8.4 Generalized Dual Simplex Algorithm with Bound-Flipping .. . . . . . 220
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 223
9 Decomposition Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 225
9.1 D-W Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 226
9.1.1 Starting-Up of D-W Decomposition .. .. . . . . . . . . . . . . . . . . . . . 231
9.2 Illustration of D-W Decomposition . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 232
9.3 Economic Interpretation of D-W Decomposition.. . . . . . . . . . . . . . . . . . 238
9.4 Benders Decomposition .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 240
9.5 Illustration of Benders Decomposition .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . 248
9.6 Dual Benders Decomposition.. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 254
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 259
10 Interior-Point Method.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 261
10.1 Karmarkar Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 262
10.1.1 Projective Transformation .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 263
10.1.2 Karmarkar Algorithm . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 264
10.1.3 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 266
10.2 Affine Interior-Point Algorithm . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 271
10.2.1 Formulation of the Algorithm .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . 272
10.2.2 Convergence and Starting-Up . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 274
10.3 Dual Affine Interior-Point Algorithm . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 276
10.4 Path-Following Interior-Point Algorithm . . . . . . .. . . . . . . . . . . . . . . . . . . . 279
10.4.1 Primal–Dual Interior-Point Algorithm.. . . . . . . . . . . . . . . . . . . . 281
10.4.2 Infeasible Primal–Dual Algorithm .. . . .. . . . . . . . . . . . . . . . . . . . 284
10.4.3 Predictor–Corrector Primal–Dual Algorithm . . . . . . . . . . . . . 286
10.4.4 Homogeneous and Self-Dual Algorithm . . . . . . . . . . . . . . . . . . 289
10.5 Notes on Interior-Point Algorithm . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 291
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 295
11 Integer Linear Programming (ILP) . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 299
11.1 Graphic Approach .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 300
11.1.1 Basic Idea Behind New ILP Solvers .. .. . . . . . . . . . . . . . . . . . . . 302
11.2 Cutting-Plane Method .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 303
11.3 Branch-and-Bound Method .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 306
11.4 Controlled-Cutting Method .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 313
11.5 Controlled-Branch Method . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 315
11.5.1 Depth-Oriented Strategy . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 316
11.5.2 Breadth-Oriented Strategy . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 319
11.6 ILP: with Reduced Simplex Framework . . . . . . . .. . . . . . . . . . . . . . . . . . . . 323
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 337
xviii Contents

Part II Advances
12 Pivot Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 341
12.1 Partial Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 342
12.2 Steepest-Edge Rule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 344
12.3 Approximate Steepest-Edge Rule . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 346
12.4 Largest-Distance Rule.. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 348
12.5 Nested Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 350
12.6 Nested Largest-Distance Rule . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 352
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 353
13 Dual Pivot Rule .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 355
13.1 Dual Steepest-Edge Rule. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 355
13.2 Approximate Dual Steepest-Edge Rule . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 359
13.3 Dual Largest-Distance Rule . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 360
13.4 Dual Nested Rule .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 362
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 363
14 Simplex Phase-I Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 365
14.1 Infeasibility-Sum Algorithm .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 365
14.2 Single-Artificial-Variable Algorithm . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 370
14.3 Perturbation of Reduced Cost. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 376
14.4 Using Most-Obtuse-Angle Column Rule . . . . . . .. . . . . . . . . . . . . . . . . . . . 380
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 383
15 Dual Simplex Phase-l Method . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 385
15.1 Dual Infeasibility-Sum Algorithm .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 385
15.2 Dual Single-Artificial-Variable Algorithm .. . . . .. . . . . . . . . . . . . . . . . . . . 389
15.3 Perturbation of the Right-Hand Side . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 395
15.4 Using Most-Obtuse-Angle Row Rule . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 398
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 401
16 Reduced Simplex Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 403
16.1 Reduced Simplex Algorithm.. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 404
16.2 Dual Reduced Simplex Algorithm . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 412
16.2.1 Dual Reduced Simplex Phase-I:
Most-Obtuse-Angle.. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 418
16.3 Perturbation Reduced Simplex Algorithm . . . . . .. . . . . . . . . . . . . . . . . . . . 422
16.4 Bisection Reduced Simplex Algorithm . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 426
16.5 Notes on Reduced Simplex Algorithm .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . 430
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 430
17 D-Reduced Simplex Method. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 433
17.1 D-Reduced Simplex Tableau . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 434
17.2 Dual D-Reduced Simplex Algorithm.. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 437
17.2.1 Dual D-Reduced Phase-I . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 442
17.3 D-Reduced Simplex Algorithm .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 445
17.3.1 D-Reduced Phase-I: Most-Obtuse-Angle Rule. . . . . . . . . . . . 452
Another Random Scribd Document
with Unrelated Content
the Bay of Biscay, which took from it eight of the convoy before they
could escape under the land. When news of this reached Brest,
Villaret was ordered to take the nine ships that alone were then
ready for sea and join the other three. This junction was made, and
on the 16th of June the twelve again fell in with the British five,
which, after temporarily withdrawing, had returned to their cruising
ground. Chase was of course given, and as two of the British sailed
very badly there was a good prospect that either they would be
taken, or the rest of the division compelled to come down to their
aid; which, with the disparity of force, ought to result in the capture
of the whole. The admiral, Cornwallis, behaved with the utmost
firmness and coolness; but with such odds and the disadvantage of
speed, no courage nor conduct can avert some disaster. The
inefficiency and bad gunnery of the French saved their enemies.
After a cannonade which lasted from nine in the forenoon of the
17th until six in the evening, one British ship had thirteen men
wounded, and Villaret abandoned the pursuit.
Five days later Bridport's fleet was sighted, and the French being
inferior stood in for their coast, intending to anchor and await action
under the island called Groix. The pursuit continued with light airs all
day and night; but at daybreak of June 23 the fastest British were
within three miles of the slowest of the French, which opened fire at
six A.M. All Villaret's signals could not bring his fleet into line, nor
induce the undoubtedly brave, but ill trained, men who commanded
the faster ships to take station for the support of the slower. A
desultory action continued till half-past eight, ending in the capture
of three French vessels, the last hauling down her flag within a mile
of Île Groix. Bridport then called his ships off. It is the opinion of
French writers, apparently shared by English critics, [101] that if he
had pursued energetically, the remainder of the enemy must either
have been taken or run ashore to avoid that fate. As it was, Villaret
was permitted to get into L'Orient without molestation, although to
do so he had to wait till the tide served. Such was the extreme
circumspection characterizing the early naval operations of the
British, until Jervis and Nelson enkindled their service with the
relentless energy of spirit inspired by Bonaparte on land. Those to
whom St. Vincent and the Nile, Algesiras and Copenhagen, have
become history, see with astonishment nine ships of capital
importance permitted to escape thus easily from fourteen; forgetting
the hold tradition has on the minds of men, and that it belongs to
genius to open the way into which others then eagerly press. How
the admiralty viewed Bridport's action may be inferred from his
retaining command of the fleet until April, 1800. The ships that
reached L'Orient had to remain till the winter, when they slipped
back two or three at a time to Brest.
The disaster at Île Groix, with some similar small misfortunes in the
Mediterranean, [102] accompanied as they were by evidences, too
plain to be any longer overlooked, of the inefficiency both of officers
and men in the French fleets, determined the government to
abandon all attempt to contest the supremacy of the sea. To this
contributed also the extreme destitution, in the dockyards, of all
sorts of stores for equipment or provision. With the English Channel
and the forests of Corsica in the hands of Great Britain, the
customary sources of supply were cut off; and moreover the
providing for the navy—not the most cherished of the national
institutions—met with the same difficulty as was experienced by
other branches of the public service in the depreciation of the
assignats, with which alone naval administrators could pay
contractors. At the end of 1795 twelve hundred francs in paper were
worth scarce twenty in gold. Having accepted naval inferiority as a
necessary condition, the Committee of Public Safety resolved to
maintain the great fleets in the ports simply as a threat to the
enemy, and to send to sea only small divisions to prey upon his
commerce and to levy tribute upon his colonies.
From the end of 1795 forward, this policy prevailed and was finally
accepted even by the emperor after several attempts, fruitless
through the incompetence of his sea-officers, to realize on the sea
that employment of great masses in which he so wonderfully
excelled on land. The reluctance of his supreme genius to accede to
this system of petty war may be accepted as testimony that on sea,
as on shore, great results can only be expected by wielding great
masses. Upon this conclusion history too has set its seal; for the
squadron and division warfare of the French navy, seconded though
it was by hosts of commerce-destroyers, public and private,
produced practically no results, had absolutely no effect upon the
issue of the war. On the other hand, that Napoleon, when convinced
that he could expect nothing decisive from his fleet, accepted the
use of it as a means of harassment or of diversion, must be received
as a weighty indication of the naval policy suited to the inferior
power. To assume a menacing attitude at many points, to give effect
to the menace by frequent and vigorous sorties, to provoke thus a
dispersion of the enemy's superior force, that he may be led to
expose detachments to attack by greater numbers,—such must be
the outline of conduct laid down for the weaker navy. But that such
a course may be really effective,—that the inferior may, as in some
of Bonaparte's wonderful campaigns, become ultimately superior,—
there must be at some fitly chosen point of the sea frontier a
concentrated body of ships; whose escape, if effected, may be the
means of inflicting a great disaster upon the enemy by crushing one
or more of the exposed fractions of his fleet. Unless there be such a
central mass, mere dissemination is purposeless. Inferiority carried
beyond a certain degree becomes impotence; nor will all the
commerce-destroyers fancy can picture restore the balance to the
nation hopelessly weaker in ships of the line-of-battle.
On land as well as on sea weakness was stamped on the military
movements of the year 1795. Its diplomatic successes were due to
the arms of 1794. Early in the year Pichegru was removed from
Holland to the upper Rhine, being succeeded by Moreau. Between
the armies commanded by these two lay that of the Sambre and
Meuse, still under Jourdan, concentrated on the left bank of the
Rhine between Dusseldorf and Coblentz. It was provided that, in
case of the three armies acting together, the chief command should
be with Pichegru. The removal of the restraint enforced by the Terror
was evidenced in these Northern forces by the number of desertions
during the severity of this unusual winter. In the army of Italy the
case was even worse; and the frightful destitution of the soldiers,
made known a year later in Bonaparte's celebrated proclamation
upon taking command, led to even greater losses from the ranks.
Toward the end of January seventeen thousand men were detached
from this army to Toulon to take part in a projected invasion of
Corsica. The expedition never sailed; but such was the horror
inspired by the sufferings they had undergone that the hardiest
soldiers forsook their colors, and of two fine divisions not ten
thousand men rejoined the army. [103]
Under such conditions military operations could not but be sluggish;
the genius of a Bonaparte was needed to supply the impetus which
the emasculated central government could no longer impart. The
great Carnot, also, had been removed from the direction of the war,
March 4, 1795, under the law prescribing periodic changes of
members in the Committee of Public Safety. The fall of Luxembourg,
already mentioned, was obtained only by the slow process of
blockade. Not till September, when the season was nearly over, did
the French armies move; Jourdan and Pichegru being then directed
to undertake a concerted invasion, from their widely separated
positions, intending to form a junction in the enemy's country.
Jourdan crossed the Rhine, advanced south as far as the Main, and
invested Mayence, the Austrians retiring before him. Pichegru
received the capitulation of Mannheim, which opened its gates upon
the threat of bombardment without receiving a shot. He showed no
vigor in following this success, and his dilatory ill-combined
movements permitted the Austrian general, Clairfayt, to concentrate
his armies in a central position between the Main and Mannheim,
parting the two French leaders. The new direction of war, either
through folly, or from sharing the growing inefficiency of the general
government, had failed to accumulate the supplies of all kinds
needed by the masses thus gathered in a wild and impoverished
region. Clairfayt combined a powerful movement against the
destitute and suffering army of Jourdan, compelling him to retreat
and recross the Rhine. Leaving a corps of observation before him
and holding Pichegru in check on the south, he then himself crossed
the river at Mayence and assaulted vigorously the lines on the left
bank with which the enemy during the past year had surrounded the
place. The blockading forces were driven back in divergent
directions; and the Austrians in increasing numbers poured into the
country west of the river, separating Jourdan from Pichegru.
It was at this moment that the National Convention was dissolving
and the Directory assuming the reins of government. Carnot became
one of its five members. Impressed by the importance of retaining
Mannheim, the loss of which was threatened by the recent reverses,
he sent urgent orders to Jourdan to move south to the support of his
colleague, using his own judgment as to the means. It was too late.
The mingled weakness and ignorance of the preceding government
had caused a destitution of means and an inferiority of force wholly
inadequate to cope with the superior strategic position of the
Austrian masses. Mannheim, which had surrendered on the 20th of
September, was regained on the 22d of November by the Austrians,
who at once re-enforced the troops on the west bank, acting against
Pichegru, while Clairfayt pushed further back the other French army.
With this success his brilliant operations closed for the year. The
weather had become excessively bad, causing great sickness; and
on the 19th of December he proposed an armistice, which Jourdan
was only too glad to accept. The Austrians, who when the year
opened were east of the Rhine, remained in force on the west bank,
holding well-advanced positions based upon Mayence and
Mannheim,—the two capital places for sustaining operations on
either side of the river.
The suspension of arms lasted until May 30, 1796, when the French,
having given the notification required by the terms of the armistice,
again crossed the river and began hostilities. But, although great
and instructive military events followed this new undertaking, the
centre of interest had by this time shifted from the north and east to
the Italian frontier of the republic, where the successes of Bonaparte
enchained it until the Peace of Campo Formio dissolved the coalition
against France. That dazzling career had begun which becomes,
thenceforth until its close, the main thread of French history, to
which other incidents have to be referred, and by following which
their mutual relations are most easily understood. We have reached
therefore the period when a naval narrative reverts naturally to the
Mediterranean; for Bonaparte's Italian campaigns profoundly
affected the political and maritime conditions in that sea. Upon it
also he next embarked for the extraordinary enterprise, condemned
by many as chimerical, and yet so signally stamped by the
characteristics of his genius, in which he first came into collision with
the Sea Power of Great Britain, destined to ruin his career, and with
the great seaman in whom that Sea Power found the highest
expression it has ever attained.
"During the year 1795," says the distinguished military historian of
these wars, "France, after a twelvemonth of victory, came near
losing all her conquests; threatened with a dangerous internal
reaction, she only with difficulty succeeded in freeing herself from
embarrassments and modifying the defects of her institutions. The
next year we shall find her launched in a yet vaster career, by the
great captain who so long presided over her destinies, who raised
her to the pinnacle of glory by his victories, and thence plunged her
to the abyss through disregard of justice and moderation." [104]
 CHAPTER VIII.

The Mediterranean and Italy.—From the Evacuation of Toulon in 1793 to

the British Withdrawal from that Sea, in 1796, and the Battle of Cape
St. Vincent, February, 1797.—Austria Forced to Make Peace.

A FTER the evacuation of Toulon, on the 19th of December, 1793,

Lord Hood had taken his fleet to Hyères Roads,—an anchorage
formed by the group of islands of the same name a few miles
east of Toulon. There he remained during the greater part of
January, revictualling his ships and considering what steps should
next be taken to assure British interests in the Mediterranean.
It is essential to a fleet, as to an army, to have always near at hand
depots of supplies, upon which it can depend to replace the
consumption of the limited stores carried on board, and to which it
can resort for refit. The way from the depots to the fleet, or the
army, technically called the communications, should be such as can
most easily be defended by the armed force itself, without impairing
its liberty in movements necessary to obtain the objects of the war.
In other words, this depot, or base, should be near the scene of
operations and, when practicable, so situated that the fleet, while
actively engaged, interposes directly or indirectly between it and the
enemy. Nearness was doubly important in the days of sailing ships,
whose movements depended upon the fickle element of the wind;
for then a ship going in for refit not only took more time to perform
the voyage, but also in returning made a heavier draft upon her
resources, and consequently could remain a shorter period at the
seat of war.
Toulon, while securely held, served as such a base; although, being
exposed to overpowering attack from the land side, its situation
called for excessive expenditure of effort. Toulon was, however, now
lost, and the British were thrown back upon Gibraltar, at the very
entrance of the Mediterranean, and nine hundred miles from the
seat of hostilities, for a secure depot in which to land stores and a
safe anchorage for transports and crippled ships-of-war.
Although the possession of a strong place, suited in itself for a base,
may decide the character of operations projected, logically the
necessary operations should first be determined, and the choice of
the base be decided by them. In the Mediterranean, as elsewhere
generally at this time, the policy of Great Britain was to control the
sea for the protection of commerce, and to sustain on shore the
continental powers in the war against France,—chiefly by money, but
also by naval co-operation when feasible. To support the land
warfare, her diplomatic negotiations strove to unite, in the general
military effort, as many as possible of the small independent states
into which Italy was divided, to promote among them unity of
action, and to foster the sense of security, in taking a decided step
against France, which they could only derive from the presence
among them of a strong power, such as Great Britain showed in her
fleet. Neither their traditions nor the character of their rulers enabled
them to combine strongly as equals; nor, as has before been said,
was there any one state so predominant as to give a nucleus around
which the others could gather. The only possible centre was Great
Britain, present as a power in her fleet. Moreover, if unable, as she
proved, to stir them up to positive, harmonious, concerted action, it
was her interest to impose upon them a benevolent neutrality
towards herself and her allies, and to deter them from inimical
measures into which they might otherwise be impelled by the
demands of the republic and their fear of the French army. Friendly
ports along the whole coast were essential for the security of her
shipping and the promotion of her commerce. Finally, it must be
added that the character of the shore between Nice and Genoa,
interposing between the French and Austrian armies, particularly
favored direct naval co-operation. The Maritime Alps and the
Apennines, coming down close to the sea, left but a single narrow,
and yet long, line of communications along the beach, traversed by
a very bad road, in many parts under fire from the sea. [105] This
condition made the armies using that line chiefly dependent for
supplies upon coasters, whose movements ships-of-war could harass
and impede,—though not, when sailing vessels, entirely stop.
Corsica, in its existing political conditions of revolt against France
under the leadership of Paoli, appeared to offer the strategic
situation which Hood was seeking. It was near and centrally placed
with reference to the probable operations on shore; San Fiorenzo
Bay, which became the chief anchorage for the fleet, being
equidistant, about one hundred miles, from Nice and from Genoa.
Leghorn also, one of the greatest depots of British trade in the
Mediterranean and the seaport of Tuscany, over which Great Britain
wished to enforce her influence, was but sixty miles from Cape Corso
at the northern extremity of the island. All the trade with northern
Italy had to pass close to Corsica, and was consequently exposed to
capture if the seaports remained in French hands; especially as
calms prevail around the island, facilitating the operations of row-
galleys and neutralizing the powers of a sailing navy. Being an
island, Corsica depended upon the control of the sea; and, though
its size and rugged surface precluded conquest, there was believed
to be a disposition to accept the protection of Great Britain as the
only means of dispossessing the French troops, who still held the
seaports of San Fiorenzo, Bastia, and Calvi.
Sir Gilbert Elliott, who had been the Civil Commissioner of Great
Britain in Toulon during the last month of its occupation, left Hyères
Bay early in January to confer with Paoli, who had proposed the
annexation of the island to the British crown. His enthusiastic
reception by the people and the assurances of the Corsican chieftain
convinced him that the measure was sincerely desired; and in
consequence of his representations, Hood, on the 24th of January,
sailed from Hyères with his whole fleet for San Fiorenzo Bay. The
weather proving very tempestuous, and the three-decked ships, of
which there were several, being ill-fitted to contend with it, the
admiral was forced to take them to Porto Ferrajo in Elba, and to
send against San Fiorenzo a detachment only. This appeared off the
place on the 7th of February; and after a series of combined
operations, in which the navy bore a very conspicuous share on
land, the French evacuated the town and works on the 19th of the
month, retreating upon Bastia. The admiral then urged an attack
upon Bastia; but the general thought he could not spare enough
men. Nelson, who had been blockading there for some time,
strongly represented the feasibility of the enterprise; and, after a
sharp altercation between the two commanders-in-chief, Hood
determined to undertake the siege with the navy and the troops who
were serving on board as marines. The landing began on the 3d of
April; but the place held out till the 21st of May, when it capitulated.
Calvi was next taken in hand, the operations beginning on the 19th
of June, and ending with its surrender on the 10th of August. The
whole island was thus freed from the presence of French troops,—a
result due almost wholly to the navy, although the army bore a
share in the operations at San Fiorenzo and Calvi. To the
determination of Hood and the ardent representations of Nelson was
due that Bastia was besieged at all; and, as thirty-five hundred
regular troops then surrendered to an outside force of fourteen
hundred seamen and marines, the opinion of Sir Gilbert Elliott that
"the blockade of the port was the chief means of reducing it," can
scarcely be disputed. At the siege of Calvi Nelson lost his right eye.
Between the siege of Bastia and that of Calvi, the General Assembly
of Corsica met, and on the 19th of June, 1794, tendered the insular
crown to the king of Great Britain. With the successful issue of the
military operations, this political act consummated the possession of
the island. But, to quote the words of Elliott, the claim to Corsica
rested upon superior force, [106] and by that only could be asserted;
and this superior force the British government failed to provide. The
love of the people for Paoli, and the period of anxiety through which
he and they had passed, caused the connection to be eagerly
desired, and accepted with demonstrations of the warmest delight;
but as security succeeded the sense of danger, the first-love
between nations so radically distinct in temperament and institutions
was followed by the symptoms attending ill-assorted unions. Still,
with prompt action and strong garrisons, the benefit of the foreign
occupation might have been manifest, dissatisfaction might have
yielded to considerations of interest, and the island been retained.
The British government acted slowly. Elliott had urged that authority
should be sent him beforehand to take over the executive functions
at once, as soon as the Act of Union passed; instead of which, Paoli
was left for four months in his old position, and developed a jealousy
of his destined successor not unprecedented in the heads of states.
This feeling, which he held with Corsican intensity, communicated
itself to his followers; and an inauspicious division of sentiment
already existed when, in October, after an interval of four months,
Elliott received his powers as viceroy. Paoli continued for a year
longer to reside in Corsica, and up to the time of his departure was a
cause of trouble to the viceroy, and so of strength to the partisans of
France, who became numerous.
During the year 1794 the superior importance of the operations on
the frontiers of Belgium and Germany, as well as in Spain, caused
the French armies of Italy and the Alps to remain quiet, after some
early successes which had placed in their hands the chief passes of
the mountains and advanced their line on the coast as far as Vado,
on the borders of Genoese territory. Early in 1795 a force of
eighteen thousand men was detached to Toulon for the invasion of
Corsica; but, although the French had in the port fifteen ships-of-
the-line, they felt that neither the admiral nor the officers possessed
the tactical skill necessary to handle the fleet in presence of an
enemy of nearly equal force, if encumbered with a large body of
transports. The case reproduced that of Conflans, when expected in
1759 to cover the French invasion of England. As a rule, in combined
military and naval expeditions the fleet and the army should start
together; but the tactical embarrassment of transports is
indisputable. If the fleet cannot encounter the enemy successfully
when not so hampered, it but encourages disaster to incur the
meeting with them in company. Not improperly, in such a case of
doubt, it was decided that the Toulon fleet should sail alone; and,
accordingly, on the 2d of March, 1795, Admiral Martin put to sea
with fifteen of the line, seven frigates, and five smaller vessels.
Despite these respectable numbers, the efficiency of the force was
poor. Out of twelve thousand officers and men on board the ships-
of-the-line, seventy-five hundred had never before been to sea; and
Martin reported that, deducting officers and petty-officers, he had
but twenty-seven hundred seamen to man the fleet. [107]
Hood had gone home the previous November, expecting to return;
and the British were now commanded by Admiral Hotham. The latter
had cruised off Toulon for three weeks in mid-winter, on account of
the indications of the French coming out, and after a succession of
most violent weather returned to San Fiorenzo Bay on the 10th of
January. He sailed again, apparently about the 22d of the month, to
cover a convoy expected from England, leaving in port the
"Berwick," seventy-four; which, through the carelessness of her
officers, had been permitted to roll her masts overboard at her
anchors. After some more hard cruising, the fleet put into Leghorn
on the 25th of February, leaving still in San Fiorenzo Bay the
"Berwick," the delay in whose repairs can only be attributed to the
penury of naval resources. The presence of the fleet in Leghorn was
probably necessary both for its own supplies and to remind the
wavering Tuscany of the power of Great Britain at her doors; but the
"Berwick" incident served powerfully to illustrate the far-reaching
effects of individual carelessness and the impolicy of exposing small
detachments not covered, directly or indirectly, by the main body. On
the 7th of March the French fleet came in sight of Cape Corso, and
almost at the same moment discovered the "Berwick," which had
only the previous day succeeded in leaving San Fiorenzo for
Leghorn. Being still crippled, she was easily overtaken, and forced to
surrender at noon.
The following day Hotham in Leghorn learned that the enemy had
sailed, and the next morning at dawn put to sea in pursuit. On the
11th the two fleets came in sight of each other, the French far south
and to windward of the British, out of gun-shot, and thus continued
during the 12th. That night one of the French seventy-fours lost a
topmast and parted company, reducing their numbers to fourteen.
The following morning the "Ça-Ira," an eighty-gun ship, ran into her
next ahead, losing both fore and main topmasts, and thus became
the source of anxiety and danger which a crippled ship ever is in the
fleet to which she belongs. As she dropped out of the line, a British
frigate ran close to her on the side encumbered with the fallen
spars, annoying her there for some time with comparative impunity,
and was then succeeded by Nelson in the "Agamemnon," who hung
about on her quarters until several ships bore down to her relief.
During the night the "Sans Culottes," of one hundred and twenty
guns, the heaviest vessel in the French fleet, dropped out of the line
to leeward, and in the morning was out of sight. The "Ça-Ira" was
taken in tow by, the "Censeur," seventy-four, and at daybreak these
two were some distance from, and to the north-east of, their fleet.
Both parties were at this time much scattered from the irregularities
of the wind; but it shifting now to the north-west favored the British,
and those nearest the separated ships speedily brought them to
action. While this partial engagement was going on, both fleets
shaped their course for the scene, the French intending to cover the
"Censeur" and "Ça-Ira" by passing between them and the British,
but owing to some misunderstanding this was not done. As the two
vans went slowly by on opposite tacks, a sharp exchange of shots
took place, in which the British, not being well closed up, were
overmatched, and two of their number suffered severely; but as the
French allowed the enemy to interpose between them and the
disabled vessels, these were forced to surrender after a very gallant
resistance, in which they lost four hundred men and were partially
dismasted. Each fleet continuing its course in opposite directions,
they soon separated and passed out of sight.
The baffling character of the wind, combined with the indifferent
sailing qualities of the British three-deckers, deprive this action of
any special tactical significance; and the very mediocre calibre of the
two admirals was not calculated to overcome the difficulties they
encountered. The numerous accidents in Martin's fleet, and the
parting of the "Sans Culottes" at such a moment, show the
indifferent character of the French captains, and probably justify the
comment that Hotham owed his success to the initiative of the
admirable officers under his command rather than to his own
capacity. "I went on board Admiral Hotham," wrote Nelson, "as soon
as our firing grew slack in the van and the 'Ça-Ira' and 'Censeur' had
struck, to propose to him leaving our two crippled ships, the two
prizes, and four frigates, and pursue the enemy; but he, much
cooler than myself, said, 'We must be contented, we have done very
well.' Admiral Goodall backed me; I got him to write to the Admiral,
but it would not do; we should have had such a day as I believe the
annals of England never produced." [108] "Admiral Hotham," says
Chevalier, [109] "showed great circumspection. He probably did not
appreciate the improvised fleets of the republic. To fight, there are
needed ships in good condition, capable seamen, skilful gunners,
and officers accustomed to order, to military dispositions, and to
squadron manœuvres. These we did not have." "The enemy," wrote
Nelson about the same fleet, a few months later, "are neither
seamen nor officers." [110] The immediate consequence of this
defeat was the abandonment of the projected expedition against
Corsica.
After the battle a gale of wind forced the British into Spezia and
caused the wreck of the "Illustrious," seventy-four, which, from
injuries received in the action, was not able to keep off shore. This,
with the capture of the "Berwick," equalized the losses of the two
fleets. The French anchored in Hyères Bay, where they were rejoined
by the "Sans Culottes," and on the 24th of March returned to Toulon.
Admiral Renaudin's arrival from Brest [111] on the 4th of April raised
the available force to nineteen or twenty ships-of-the-line; to which
the enemy could now oppose only thirteen of their own and two
Neapolitan. But, although the British were very destitute of material
for necessary equipment and repair, [112] the superiority of
organization, discipline, and officer-like training, allowed them little
real cause for anxiety. In the month of May a Jacobin outbreak
occurred in Toulon. The government had ordered Martin to take
advantage of Renaudin's junction, and again to seek the enemy
while thus superior. The mob raised the cry that as soon as the fleet
sailed the enemies of the Revolution would enter the town and
massacre the patriots. The seamen, except those belonging to the
Brest squadron, left their vessels, under pretext of deliberating on
the dangers of their country, and took part in all the street
demonstrations. An expedition was even set on foot against
Marseille; but the central government was getting stronger, or rather
anarchy was becoming wearisome, and after a paltry engagement
the Toulonese fell back upon their city and submitted. A great
number of seamen, however, had deserted; and the necessity of
recovering them delayed the departure of the fleet, which finally put
to sea on the 7th of June with seventeen ships-of-the-line. The time
for a favorable sortie was, however, past; the British having at last
received a large re-enforcement.
After partially refitting in Spezia, Hotham took his fleet to San
Fiorenzo, arriving there on the 30th of March. On the 17th of April
he sailed again for Minorca, where he hoped to meet a re-
enforcement and a much-needed convoy; but a succession of
westerly winds caused the fleet to lose ground, instead of gaining,
until the 27th, when it unexpectedly and most fortunately met a
body of store-ships from Gibraltar. With these the admiral at once
bore up for Leghorn, arriving there next day. Had the French in their
own dockyards refitted as rapidly as the British did in a foreign port,
and gone to sea with the increase of strength that Renaudin
brought, they should have intercepted this important convoy, if they
did not bring the fleet to action. The correspondence of Nelson, still
a simple captain, testifies continually, in his own vivid style, to the
critical state of the campaign at this moment.
On the 8th of May Hotham sailed again for Minorca, and continued
to cruise off that island until the 14th of June, when Rear-admiral
Robert Mann joined him with a re-enforcement from England of nine
[113] ships-of-the-line, making the British superior in number as in
quality to the enemy. As a convoy was also expected, the fleet kept
the same station until it arrived on the 22d, when the whole body
sailed for San Fiorenzo, anchoring there June 29. On the 4th of July
Hotham sent Nelson with his own ship and some smaller vessels to
co-operate on the Riviera of Genoa with the Austrian advance
against the French; but the detachment fell in with Martin's fleet,
and had to return to San Fiorenzo, being, says Nelson, "hard
pressed" by the enemy, who pursued until they saw the British fleet
at anchor. The latter weighed as soon as possible, but did not come
up with the French until the 14th, near the Hyères Islands, where a
trifling brush took place, resulting in the capture of the "Alcide,"
seventy-four, which immediately after surrendering caught fire and
blew up. This small affair was the last in which Admiral Hotham was
directly concerned. Hood had been definitively relieved of the
Mediterranean command before he could sail from England on his
return; and Hotham, weary of a burden to which he felt himself
unfitted and had proved himself unequal, applied for relief, and
struck his flag on the 1st of November, 1795.
Prior to the naval brush off Hyères, little had been done in Italy by
the armies on either side. The French, whose forces were far inferior
to those of the allies and in a state of great destitution, were
compelled to stand on the defensive over a very long line, of which
the advanced post on the sea was at Vado. The Austrians had
imposed upon the Sardinians their own plan of campaign, which was
to strike the extreme right of the French at Vado, and then drive
them back along the Corniche, while the Sardinians attacked on the
other flank through the passes of the Apennines up to the Col di
Tende, where the Maritime Alps begin. The British fleet was to co-
operate with the movement on the side of the sea,—a consideration
which had much to do with determining the Austrian plan. By
occupying Vado and the Riviera east of it, the coasting trade
heretofore carried on from the ports of Genoa and Tuscany to
southern France would be stopped; a matter of great consequence
to the republic, as those departments only raised grain for three
months' consumption and depended for the rest upon that which
came from Barbary and Italy by way of Tuscany. With a British
squadron at Vado, the populace of Provence, the navy at Toulon,
and the army of Italy, would have to be supplied from the north of
France by bad inland roads. [114] Vado Bay was, moreover, the best
anchorage between Nice and Genoa. Devins, the Austrian general,
began his movement on the 13th of June, passing through Genoese
territory against the protest of that neutral government. The French
resisted sturdily; but the odds were too great, and by the end of
June they had fallen back to a line extending from Borghetto on the
sea to Ormea in the mountains, abandoning Vado and the
intermediate coast towns. Devins now called on Hotham for naval
support, and the British admiral detached Nelson, as has already
been seen, to give it; but the encounter with the French fleet
delaying his arrival for some days served Devins with a pretext for
inaction. He employed the mean time in fortifying his position and
improving the roads in his rear.
 Northern Italy and Corsica.
Click here for larger map

On the 17th of July Nelson reached Genoa, took on board the British
minister to that republic, and went to Vado, anchoring there on the
21st. He immediately conferred with Devins, who gave it as his
opinion that the Austrians could not advance further until the French
were compelled to retire by want of provisions,—a result he
expected from the action of the navy. Nelson explained to the
minister that the food supplies being carried in neutral vessels he
was powerless to stop them; for he had stringent orders from the
admiralty that no legal measures should be instituted against such,
when arrested, until full particulars in each case had been sent to
England and instructions had come back what to do. Meanwhile the
cargoes, being perishable, would be spoiled. Such utterly
inconsequent proceedings, though highly characteristic of the
military action of cabinets, were most exasperating to a man of
Nelson's temper, always prone to cut knots; but, as he was
pecuniarily answerable and a poor man, he could not disregard them
on his sole authority. The British ministers to Genoa and Sardinia
both concurred in advising him to carry out Devins's wishes; and
then Nelson, with the fearlessness of responsibility he always
showed,—and sometimes out of, as well as in, season,—issued
orders to his frigates to take every vessel bound to France or to
ports within the French lines, to whatsoever nation it might belong.
At the same time the general also sent out small cruisers; and it is
roundly asserted by Jomini that he laid a tax on all coasting vessels
brought in by them, forcing them to pay for a passport and
appropriating the proceeds to himself. The utmost diligence,
however, could not insure the interruption of a traffic carried on by
very small vessels, having to make but short night runs close along a
coast; [115] and Nelson's orders, while doubtless productive of some
results in stopping large vessels from Tuscany and Algiers, could not
prevent supplies getting in by Genoese coasters. The difficulty was
increased by the fact that a number of the ports in rear of the
French positions belonged to Genoa, and the inhabitants depended
upon her for food. To turn a deaf ear to their cries of suffering, and
to prevent supplies landed there being used by the French, were
alike impossible.
These considerations should have led Devins to regard naval co-
operation simply as the efforts of a light force, competent to harass,
not to destroy, the enemy's communications; and should have
induced him to force his own way, by early and vigorous action,
while the French were inferior in numbers and before their positions
became too strong. Instead of this, he used the Sardinians and
British, not as allies, but as excuses for not moving; action of one
kind or another on their part was necessary before he could
advance. Thus the summer slipped away, the French busily
strengthening their lines and bringing up the army of the Pyrenees
to re-enforce that of Italy, after peace was made with Spain. Nelson
was full of projects to embark a corps of Austrians, who should seize
and occupy a coast position in rear of the French; but Devins only
played with a proposition of which Jomini speaks with scant respect,
and which was certainly open to the objection that, if carried out
with adequate force, it divided the Austrian army. Nelson, it is true,
guaranteed its retreat to his ships in case of need; but the Austrian,
not unnaturally, preferred a less uncertain line. The British captain,
who had at first felt respect for Devins's abilities, wrote on the 17th
of September that it had for some time appeared to him "that the
general intended to go no further than his present position, and
meant to lay the miscarriage of his enterprise against Nice to the
non-co-operation of the British fleet and the Sardinian army." [116]
Whatever his purpose, Devins did not advance, but quietly awaited
the French attack. Early in November a severe illness caused him to
give up the command; and on the 23d of the month the enemy,
under General Schérer, assaulted vigorously the centre of the allied
position, where the Sardinians touched the Austrians. This point had
always been weak, and after a short struggle was forced. The
original intention of Schérer had been to turn, after piercing the
centre, against the Sardinians, as Bonaparte did in the following
year; but a very heavy fall of snow in the mountains decided him to
swing round to the right, drive the Austrians back upon the coast,
and, if possible, anticipate them upon their line of retreat to the
eastward. In this he was not entirely successful; but the imminent
danger forced the enemy to abandon all their line of works and fall
back precipitately, with a loss of seven thousand men, killed,
wounded and prisoners, besides their depots at Loano, Vado, and
Savona. This action, which was a brilliant and decisive victory, is
known as the battle of Loano. By the first of December the Austrians
had recrossed the Apennines and were again in the positions from
which they had set out the previous June.
The military plan and execution of the campaign by the land forces
scarcely falls within the scope of the present work. An advance by
the armies of a coalition, whose respective troops and lines of
operation are separated by a chain of mountains, even of the height
of the Apennines, with difficult communication across them, does
not present a combination promising mutual support and probable
success. To this disadvantage is to be added that of the long and
narrow line by which all the Austrian supplies had to be forwarded;
and which any successful advance would make yet longer and more
difficult. On the other hand, the Austrians counted much upon the
help that could be extended by the navy of Great Britain, in whose
interest the occupation of the shore line was undertaken. [117] By
their advance they controlled Vado Bay,—the best anchorage
between Villefranche (Nice) and Genoa; and their presence imposed
a restraint upon the latter republic, whose attitude was largely
determined by the comparative forces of the belligerents. Did then
the British navy, under these circumstances, do all that it could have
done to insure the success of the common cause?
The answer can scarcely be yes. Nelson, indeed, exerted himself
with the energy that never failed him; but his correspondence shows
that he did not think the force assigned him equal to its task. Could
it then have been increased? The answer again is scarcely doubtful.
The British fleet of the line was slightly superior in numbers to the
French at Toulon, and far superior in the quality of its officers and
men. It was doubtless embarrassed by numerous duties,—by those
conflicting interests whose divergence imposes the great test of
capacity upon a general officer. The French fleet at Toulon, co-
operation with the Austrian advance, the protection of trade, the
covering of Corsica, the political interests involved in controlling the
action of the small Italian states,—all these cares fell upon the
British admiral. Of these, the French fleet was the most important;
but all other interests of Britain and her allies were better served by
co-operation with the Austrian advance,—by victory in the field,—
than by any dissemination of force for other purposes. A decisive
Imperial success would have determined the policy of every state in
the western Mediterranean and closed every port to French cruisers.
In short, offensive action, and not the merely defensive attitude
maintained through the campaign, was here clearly indicated.
Nelson intimates that the only course by which the navy could
practically intercept the French communications was to enter the
coasting ports and destroy the coasters. These little vessels defied
detection on their voyages; only by chasing them into their nests
could their wings be clipped. "A few days ago," he writes, "I scoured
the coast between Monaco and Borghetto so completely, that
although I was only able to take one ship loaded with corn, yet I
forced the others into the Bay of Alassio, where they are so
completely under the protection of formidable batteries that not less
than three sail-of-the-line could attempt to take or destroy them.
The number of vessels loaded and unloaded at those places is near
one hundred, the greater part loaded with stores and corn for
France." [118] Here was the strategic direction to be given to the
British navy, after providing for the watch off Toulon. "You will now,"
wrote Nelson five years later to Lord Keith, "bear me out in my
assertion, when I say that the British fleet could have prevented the
invasion of Italy; and if our friend Hotham had kept his fleet on that
coast, I assert, and you will agree with me, no army from France
could have been furnished with stores or provisions; even men could
not have marched." [119] If the fleet proved unequal to this task,
final condemnation was passed on the allied plan of campaign,
which was not in its conception characterized by sound military
judgment. But Admiral Hotham, as Nelson said, had "no head for
enterprise, perfectly satisfied that each month passes without any
losses on our side." [120] Nelson never had under his orders any
other ship-of-the-line than his own "Agamemnon;" and at the time
of the decisive battle of Loano all his little squadron had been taken
away except two, so that French gunboats harassed with impunity
the left flank of the Austrians. [121] He himself at that critical
moment had to remain in Genoa with the "Agamemnon," at the
request of the Imperial minister, to prevent the crew of a French
frigate then in port, supported as it would have been by French
partisans, from seizing Voltri,—an important point upon the line of
retreat of the Austrians, where a few resolute men could have
stopped them until the pursuing army came up. To him alone was
therefore attributed the escape of several thousand Imperial
soldiers, among whom was the commander-in-chief himself. [122]
Prior to the battle Genoa had permitted French intrigues and armed
enterprises of this character to be concerted, almost openly, in her
territories. This she would not have dared to do, had the British navy
been present in force on the coast, acting under such a commander
as Nelson; for the probabilities of final success would have been with
the allies. In short, this campaign of the British fleet contributes
another to the numerous lessons of history, upon the importance of
having sufficient force at the decisive point and taking the offensive.
It may be added that Hotham could better have spared ships to
Nelson, if he had not thrown away his two opportunities of beating
the Toulon fleet.
While Nelson was co-operating with the Austrians as far as his force
admitted, the British fleet was generally cruising off Toulon,
returning from time to time to San Fiorenzo or Leghorn for refit and
stores. It was in this latter part of the year 1795 that the Directory,
as will be remembered, decided to abandon the policy of fleet-
fighting and to enter upon that of commerce-destroying, directed
against exposed colonies of the enemy as well as against his trade
afloat. Two squadrons, numbering in all seven ships-of-the-line and
eight smaller vessels, were ordered fitted out at Toulon, which was
with difficulty done for want of seamen. Since the action of July off
Hyères, nearly all the sailors in Martin's fleet had deserted, disgusted
with the bad food, scanty clothing, and constant disaster that were
their portion. Enough, however, were at last gathered to man the
ships selected; and on the 14th of September six of the line and
three frigates got away under the command of Captain Richery. It
does not appear whether the British fleet was then at sea or at San
Fiorenzo; but in either case it was at this port, and not until
September 22d, that Hotham learned their escape. On the 5th of
October this leisurely commander-in-chief sent Admiral Mann with
six of the line in pursuit; but the French, having so great a start,
could not be overtaken. They passed the Straits of Gibraltar early in
October, bound to the British possessions in North America. On the
7th of the month, when a hundred and fifty miles west of Gibraltar,
they fell in with an enemy's convoy of thirty-one merchant ships
from the Levant, under the protection of three seventy-fours.
Richery succeeded in capturing one of the latter, which had lost a
topmast, and all the merchant ships except one. Having so valuable
a booty, he decided to escort it into Cadiz, where he anchored on
the 13th and was soon after found by Mann, whose arrival
prevented his departure to fulfil his original mission. At about the
same time some French frigates in the Atlantic took eighteen ships
out of a Jamaica convoy. The other Toulon division, of one ship-of-
the-line and six smaller vessels, cruised in the Levant; and, having
made a number of prizes, returned safely to Toulon. Its commander,
Captain Ganteaume, though undistinguished by any great
achievements, was throughout his career remarkably fortunate in
escaping the search of an enemy. It was he who commanded the
flotilla on board which Bonaparte stole unseen through all the British
cruisers, on his return from Egypt to France in 1799.
These results, coinciding so closely with the adoption of the new
policy of commerce-destroying, confirmed the government in favor
of this course, to which the French have always been strongly
disposed. They hoped from it, to use the words of a representative
in the Convention, "to force the English to a shameful bankruptcy;"
what they obtained was the demoralization of their navy, the loss of
the control of the sea and of their own external commerce, finally
Napoleon's Continental System and the fall of the Empire.
The battle of Loano, decisive of the campaign of 1795, is yet more
distinguished as marking the entrance upon the scene of two of the
most remarkable figures in the war of the French Revolution. During
the week after it was fought, Admiral Sir John Jervis, better known
by his later title of Earl St. Vincent, arrived at San Fiorenzo, as the
regular successor to Hood in the Mediterranean. During the winter
Napoleon Bonaparte was chosen by the Directory to relieve Schérer
in command of the Army of Italy.
The career and character of the youthful republican general are too
well known, have been too often described, to be attempted by the
author, from whose immediate theme, moreover, they stand apart.
The personality of the already aged admiral, whose iron hands
stamped his own image on the British navy and fashioned it into the
splendid instrument with which the triumphs of Nelson were won, is,
on the contrary, familiar to few except the students of naval history.
Born in 1734, Sir John Jervis, when he assumed command of the
Mediterranean fleet in his sixty-second year, had had no opportunity
of distinguishing himself in the eyes of the world outside of the
service to which he belonged. With the members of that service,
however, he had long been a marked man. The child of a poor
though well-born family, he had in early life, under the pressure of
poverty, required of himself the same stern discipline and submission
to the duty of the moment which he afterwards so rigorously
exacted of others. Grave and unbending in his official relations,
immovable as a rock when his determination was once formed,
unrelenting almost to mercilessness in suppressing insubordination,
then rife throughout the British navy, he had the high-bred polish of
a man used to good society, and his demeanor was courteous, and,
when occasion demanded, even courtly. A traveller by land as well
as by sea, a constant and judicious reader, in a period when such
habits were rarer than now among seamen, he was well informed in
matters other than those relating merely to his profession. Of the
latter, however, he was a master. His ship was the model of the
British fleet during the American Revolution, and his high reputation
drew to her quarter-deck youths from the best families of England,
when they could obtain interest to get there. Yet no man was ever
less swayed, in an age when social and political influence counted
for so much, by any extrinsic claims of that character. Personal merit
first, after that a family claim upon the navy, that a father or brother
had given his life for the service,—nay, the very friendlessness of a
deserving man,—such were the considerations that determined him
in the use of patronage at his own disposal.
Yet, with all these strong attributes, capable too of a tenderness
which could mourn long and deeply the loss of a valued comrade,
the rule of Jervis was one of fear rather than love. It is impossible to
criticise adversely measures whose extreme severity was justified, if
not imperatively demanded, by the appalling crisis of the mutinies of
1797; impossible to withhold admiration, not unmingled with awe,
from the impressive figure of the chief who stood unmoved and
unyielding amid the smothered discontent, revolt threatening from
below, the enemy's coast in sight from the deck, knowing that in
every other fleet the crews had taken the ships from their officers,
but determined it should not be so in the Mediterranean. Yet
admiration is qualified by the feeling that to the ruthless temper of
the man the position was not wholly displeasing,—that he was in his
natural element when crushing opposition. A captain, who with great
personal courage had quelled a rising in his ship, dragging the
ringleaders with his own hands from among their followers,
interceded on behalf of one of the condemned because he had
previously borne a good character. "I am glad of it," replied Jervis;
"hitherto we have been hanging scoundrels. Now men will know that
no good character will atone for the crime of mutiny." In lesser
matters, also, his tendency was to exaggerate restraint as well as
punishment. "Where I would take a penknife," said Nelson, "Lord St.
Vincent takes a hatchet."
With such characteristics, accompanied though they were by
resolution and high professional accomplishments, it is not to be
expected that the fire of genius will be found. Though not an
ungenerous man, Lord St. Vincent lacked the sympathetic qualities
that made Nelson at once so lovable and so great a leader of men.
Escaping the erratic temper and the foibles of his great successor,
upon whose career these defects have left marks ever to be
regretted by those who love his memory, Jervis fell short too of the
inspiration, of the ardor, which in moments of difficulty lifted Nelson
far above the common plane of mankind, and have stamped his
actions with the seal of genius. But after Nelson, Jervis, though of a
different order, stands first among British commanders-in-chief. For
inspiration he had a cool, sound, and rapid professional judgment;
for ardor, a steady, unflinching determination to succeed; and these,
joined to a perfect fearlessness of responsibility such as Nelson also
showed, have won for him a place in the first rank of those
chieftains, whether sea or land, who have not received the
exceptional endowments of Nature's favorites. In the one general
action which Fortune permitted to him, the battle of Cape St.
Vincent, he illustrated these traits to a high degree; as Nelson then
also showed that faculty of quick appreciation and instant action, in
which all the processes of thought and will blend into one
overpowering conviction and impulse that lesser men never know.
Whether we consider the vastly superior numbers then deliberately
engaged, the tactics of the admiral on the battle-field, or his
appreciation of the critical position in which Great Britain then stood,
Sir John Jervis's conduct on that occasion must make the battle of
Cape St. Vincent ever illustrious among the most brilliant sea-fights
of all ages.
To these powerful elements of his nature, Jervis added a capacity for
comprehensive and minute attention to the details of discipline,
order and economy, without which mere severity would become
aimless and productive of none but bad results. He was fortunate in
finding among the Mediterranean captains an unusual number of
men of consummate seamanship, energy and resources, in all the
vigor of a prime still youthful, who were only waiting for a master-
hand to combine and give direction to their abilities. With such a
head and with such subordinates, the British Mediterranean fleet
soon became a model of efficiency and spirit, which was probably
never equalled in the days of sailing ships. Nelson so considered it;
and the old admiral himself bewailed its memory several years later,
when commanding the Channel fleet, and complained testily of the
"old women in the guise of young men," whom he found in charge
of ships off Brest. As an administrator, when First Lord, the economy
of Jervis became exaggerated into parsimony, and his experience of
the frauds connected with the dockyards of the day led him into a
crusade against them, which was both well meant and necessary,
but particularly ill-timed. It has consequently left a stigma of failure
upon his administration, which is due, however, not to his executive
inefficiency, but to a misapprehension of the political signs of the
times. Absorbed in reform, and for it desiring quiet, he saw only
peace while the dark clouds of war were gathering thick on the
horizon. Therefore the British navy, well-worn by the first war, was
not ready for that which followed it in 1803.
Jervis's arrival in the Mediterranean was too late to remedy the
impending evils. It was a singular misfortune for Great Britain, that
the interregnum between two such able men as Hood and Jervis
should have coincided with the determination of the French to try
the chance of battle with their Mediterranean fleet, and that the
opportunities they lost should have fallen to so sluggish and cautious
an admiral as Hotham. "To say how much we wanted Lord Hood on
the 13th of July," wrote Nelson, "is to say, Will you have all the
French fleet, or no action?" [123] Accepting this opinion in the light of
Nelson's subsequent achievements, it may be permitted to think
that, if not all the fleet, so many ships would have fallen as to have
prevented the sailing of Richery's squadron and the consequent
necessary detachment of Admiral Mann; while the loss of seamen
captured would have seriously crippled the operations of the flotilla,
which from Toulon supplied the Army of Italy with ammunition,
artillery and stores. Light guns on mountain carriages could be
carried along the Corniche; but at the opening of Bonaparte's
operations all heavy guns, and artillery outfits of all kinds, had to be
taken by sea from Nice to Savona. [124] The demands of this flotilla
necessitated the laying up of the fleet, [125]—a matter of less
consequence as the determination to resort to commerce-destroying
had then been reached. The French navy therefore became, through
the flotilla, a very important part of Bonaparte's communications;
and it has already been pointed out that sailing ships could not
break up, though they might much disturb, the voyages of the
smaller vessels employed on a difficult coast, with batteries under
which to take refuge.
After the battle of Loano, Nelson, whose occupation on the Gulf of
Genoa was for the time over, went to Leghorn to refit his ship, then
nearly three years in commission. Not till January 19,1796, did he
join Jervis, who since his arrival on the station had, for the most
part, remained in San Fiorenzo organizing his fleet. The new admiral
showed him the same confidence as his predecessors, and sent him
at once to his old station, with a light division, to prevent any small
number of men making a descent upon Italy. A predominant idea,
one might almost call it a fad, in Nelson's mind, was the landing of a
body of men from ships in rear of the enemy. As has been seen, he
was forward to recommend such an attempt to Devins, promising to
support it with his squadron; and the intelligence concerning flat-
boats and gunboats prepared in the French ports suggested nothing
to his mind so much as transporting troops to Tuscany, in rear of the
Austrians, while the main French army operated in their front. Like
Bonaparte, Nelson recognized the resources which the plains of
Piedmont, Lombardy and Tuscany would offer to the needy enemy.
He called them a gold mine; but he did not understand the
weakness of the French in seamen, nor realize the improbability of
Bonaparte's attempting such a use of his troops as would put them
far out of mutual support and away from his own control. Certainly
no indication of such a purpose is to be found in his correspondence
or in the instructions of the Directory to him. On the contrary, he
had strongly advised against a pet project of the Committee of
Public Safety, early in 1795, to land an expedition in the papal
states,—unless with control of the sea. [126]
Had the Austrians again advanced to the sea and occupied Vado,
Jervis would undoubtedly have supported them and harassed the
French to a very important extent. Nelson gave express assurances
on that point. [127] Bonaparte, however, allowed him no opportunity.
Leaving Paris on the 14th of March, 1796, the young general
reached Nice on the 27th of the month. On the 5th of April he
moved his headquarters to Albenga, and on the 9th to Savona. On
the 10th Beaulieu, the new Austrian general, began to move his left
wing by the pass of La Bochetta, his right by that of Montenotte; the
junction to be formed at Savona. Quick as lightning, Bonaparte
struck at once where the Austrian right touched the Sardinian left.
Blow followed blow upon the centre of the allies, and after six days'
fighting their armies were definitively separated. Driving the
Sardinians before him in unremitting pursuit, Bonaparte on the 28th
granted an armistice, by which three of the principal fortresses of
Piedmont were put into his possession and plenipotentiaries
dispatched to Paris to treat for peace. This was concluded and
signed on the 15th of May. Sardinia abandoned the coalition,
surrendered the counties of Savoy and Nice, and yielded other
conditions favorable to France,—particularly in the boundary lines
traced on the crests of the mountains, where the commanding
military positions were given to the republic. Thus the gates of Italy
were forced; and Austria, stripped of her ally on shore and cut off
from the British at sea, alone confronted Bonaparte.
The French were now in the plains of Piedmont, with Lombardy
before them. Beaulieu, expecting an advance against Milan by the
north bank of the Po, had withdrawn across that river, intending to
dispute its passage. If forced, he would cover Milan by falling back
successively upon the lines of the Sesia and the Ticino, tributaries of
the main stream. Bonaparte, however, was not the man to attack an
enemy in front and force him back along his natural line of retreat to
his proper base. Weighing accurately the political and military
conditions of the peninsula, he had fixed his eye upon the line of the
Adige as that which he wished to reach and hold, and which, under
all the circumstances, he believed he could master. The Adige flows
from the Tyrol south along the east shore of the Lake of Garda, then
turns to the eastward and enters the Adriatic between the Po and
Venice. Occupying it, the French army would cover all the valleys of
the Po, lay tribute upon their resources as well as upon those of the
small states south of the river, interpose between Austria and
southern Italy, and isolate Mantua, the enemy's great stronghold.
Making, therefore, a feint of following Beaulieu by the methodical
front attacks expected by him, Bonaparte pushed his main force
stealthily along the south bank of the Po. On the 7th of May the
advance-guard reached Piacenza, and crossed at once by boats. On
the 9th a bridge was completed over the river, which at this point is
fifteen hundred feet wide and very rapid. Beaulieu's intended
positions on the Sesia and Ticino were thus turned, and the
Austrians necessarily fell back to the line of the Adda. On the 10th of
May, just one month after Beaulieu began his forward movements,
the bridge of Lodi, over the Adda, was carried; and the Austrians
again fell back to the Mincio, the outlet of the Lake of Garda,
uncovering Milan. On the 15th Bonaparte entered Milan in triumph.
Here he paused for ten days, and, after quitting the place, had to
return to punish a revolt which broke out among the people; but on
the 30th of May the French crossed the Mincio, the Austrians
retreating northward toward the Tyrol, along the east shore of the
Lake of Garda.
This retrograde movement left Mantua to itself. On the 3d of June
Bonaparte's headquarters were at Verona,—a strongly fortified place
bestriding the Adige, thus insuring an easy transit to either side of
the river, and which derives further strategic importance from its
topographical position. A number of spurs run south from the Tyrol
along the Lake of Garda and fall into the plain at Verona, which thus
stands at the foot of the valleys formed by them. On either side of
this cluster of spurs lie the valleys of the Adige and the Brenta, the
two probable lines by which an Austrian attack would come. Verona,
therefore, was a central point with reference to any offensive
movements of the enemy, and became the pivot upon which
Bonaparte's strategy hinged. On the 4th of June Mantua was
blockaded. Having now compassed his first objective, Bonaparte
passed temporarily from the offensive to the defensive, ceased his
advance, and occupied himself with assuring the line of the Adige
and pressing the siege of Mantua.
There remained only to realize the political advantages gained by his
wonderful successes. The Duke of Parma had entered into a
convention on the 9th of May, followed in the same course on the
17th by the Duke of Modena. On the 5th of June the Court of
Naples, startled out of its dream of security, signed an armistice,
withdrawing its troops from the coalition and its ships from the
British fleet,—a precipitate abandonment of the common cause as ill-
judged as it was cowardly. At that very moment the French leader
was writing, "I see but one means not to be beaten in the autumn;
and that is, so to arrange matters that we shall not be obliged to
advance into southern Italy." [128] The Pope still holding out,
Bonaparte improved the time which the Austrians must need to
prepare a new movement, by marching into the papal states a corps
under Augereau, whom he followed in person. On the 19th Bologna
was reached, and on the 24th the Pope signed an armistice.
Coincidently with this advance, it was felt safe and opportune to
send into Tuscany a division taken from the corps occupying
Piedmont. This detachment entered Leghorn on the 28th of June,
occupied the port despite the neutrality of Tuscany, drove out and
broke up the great British commercial and naval interests centred
there, and obtained a secure base for the intended attempt upon
Corsica.
The failure of the Austrians to reach the coast, and their subsequent
retreat, of course put an end to any direct co-operation between
them and the British fleet. Jervis was forced to confine himself to
watching the Toulon ships,—an operation conducted in the same
spirit and on the same system which he afterwards imparted to the
Brest blockade, and generally to that of all hostile arsenals. For over
six months, from the beginning of April to the middle of October, he
cruised with fifteen sail-of-the-line off the port; the heavy ships
remaining some distance from it, but near enough to support a light
division of three seventy-fours, which kept just out of range of the
batteries, about two miles from the entrance. By unremitting care
and foresight, the ships on this arduous service were provisioned,
watered and repaired on the spot, without going into harbor. Nelson,
as the year before, was actively employed in the Gulf of Genoa,
harassing the coast communications, and was on one occasion
fortunate enough to capture a convoy with guns and entrenching
tools for the siege of Mantua. In the Adriatic, a few frigates and a
flotilla of small vessels were engaged in protecting the Austrian
communications by way of Trieste. Admiral Mann, with seven ships-
of-the-line, was still off Cadiz, in the station assigned him by Hotham
to watch Richery. Besides these strictly military operations, ships
were called for in every direction to convoy trade, to cover the
passage of storeships, and generally to keep the sea safe for