Scribd
Upload
44 views16 pages

CS403 Algorithm Design and Analysis: Course Content and Introduction

This document outlines the content and structure of the CS403 Algorithm Design and Analysis course, including an introduction to algorithm design paradigms and analysis of efficiency, as well as representative algorithm problems such as interval scheduling, weighted interval scheduling, bipartite matching, and independent set. It also provides an example of using the Gale-Shapley algorithm to find a stable matching and proves some of its properties.

Uploaded by

divyanshuvarma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
44 views16 pages

CS403 Algorithm Design and Analysis: Course Content and Introduction

This document outlines the content and structure of the CS403 Algorithm Design and Analysis course, including an introduction to algorithm design paradigms and analysis of efficiency, as well as representative algorithm problems such as interval scheduling, weighted interval scheduling, bipartite matching, and independent set. It also provides an example of using the Gale-Shapley algorithm to find a stable matching and proves some of its properties.

Uploaded by

divyanshuvarma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

CS403 Algorithm Design and Analysis:

Course Content and Introduction

Satyajit Thakor
IIT Mandi

16 February, 2017
Course content

I In previous courses, you have studied some algorithms, their


implementation and the problem they are used to solve.
I In this course, the order is changed:

Problem Algorithm Design Analysis implementation


I Objective: For a given problem
I How to design an algorithm?
I How to analyze its performance?
I Implementation for hands-on experience (lab component).
I Content: design paradigms, analysis of efficiency, complexity
classes.
Textbook, attendance & evaluation

I Textbook: Algorithm Design by Jon Kleinberg and Eva Tardos,


Pearson, 2006.

I Institute attendance policy: 70% to qualify for final exam

I Assessment:
Theory (60%)
I 18% Quiz 1 (approx.)
I 18% Quiz 2 (approx.)
I 24% Final Exam (approx.)
Practical (40%)
I Lab experiments/reports/assignments/exams/viva
A word of caution

I To learn the ideas well, it is important to solve as many


problems as possible.

I Slides are not (sufficient) reading material


- They just provide a map and you cannot make a journey on a
map!
- More information will be provided during lectures and
students should also refer books to learn the subject.
Moodle

I Important: regularly visit Moodle


I Syllabus, lectures, slides, assignments
I It is also an important platform for our discussions.
I Post your queries on Moodle and I will respond to it so that
everybody will learn.
Stable matching

I Problem: Design a university admission (or industry


recruitment) process that is self-enforcing.
I Self-enforcing agreement - mutually beneficial agreement
I Students (interns) with preference for universities (firms)
I universities (firms) with preference for students (interns)

I A simpler problem: devise a system by which each of n men and


n women can end up getting married given their preferences.
I M = {m1 , . . . , mn } - set of men
I W = {w1 , . . . , wn } - set of women
I M W = {(m, w) : m M, w W } - set of all ordered pairs
Problem

I A matching S is a set of ordered pairs, such that each member


of M and each member of W appears in at most one pair in S.
I A perfect matching S 0 is a matching such that each member of
M and each member of W appears in exactly one pair in S 0 .
I Preferences: each person ranks persons of other gender.
I For example, m prefers w to w0 if m ranks w higher than w0 .
I Simplifying assumption: Ties are not allowed in the ranking.
I That is, no person is allowed to have equal preference for two or
more persons of the opposite gender.
I That is, each preference list constitute a strictly partial order.
I Can there be any issue with a perfect matching S?
Problem

I Can there be any issue with a perfect matching S?


I Consider the situation: two pairs (m, w) and (m0 , w0 ) in S such
that m prefers w0 to w, and w0 prefers m to m0 . Now what ?
I (m, w0 ) will run away! - the set of marriages S is not
self-enforcing.
I A pair (m, w0 ) is an instability with respect to S: (m, w0 ) does
not belong to S, but each of m and w0 prefers the other to their
partner in S.
I A matching S is stable if (i) it is perfect, and (ii) there is no
instability with respect to S.
I Q1: Does there exist a stable matching for every set of
preference lists?
I Q2: Given a set of preference lists, can we efficiently construct a
stable matching if there is one?
Examples

I Example 1
m prefers w to w0 , m0 prefers w to w0
w prefers m to m0 , w0 prefers m to m0

I {(m, w), (m0 , w0 )} is a unique stable matching


I {(m0 , w), (m, w0 )} is perfect but not stable.

I Example 2
m prefers w to w0 , m0 prefers w0 to w
w prefers m0 to m, w0 prefers m to m0

I {(m, w), (m0 , w0 )} is a stable matching (in favour of men)


I {(m0 , w), (m, w0 )} is also a stable matching (in favour of women)
Algorithm design
I The intuitive procedure: Gale-Shapley algorithm
Analysis

I Observation 1: w remains engaged from the point at which she


receives her first proposal; and the sequence of partners to
which she is engaged gets better and better (in terms of her
preference list).
I Observation 2: The sequence of women to whom m proposes
gets worse and worse (in terms of his preference list).

I Proposition 1: The G-S algorithm terminates after at most n2


iterations of the While loop.
I Proof [Hint: What should you be counting at each iteration?]
Analysis

I Proposition 2: If m is free at some point in the execution of the


algorithm, then there is a woman to whom he has not yet
proposed.
I Proof

I Proposition 3: The set S returned at termination is a perfect


matching.
I Proof

I Theorem 1: Consider an execution of the G-S algorithm that


returns a set of pairs S. The set S is a stable matching.
I Proof
The approach

For a given practical problem:


I formulate the problem with enough mathematical precision that
we can ask a concrete question and start thinking about
algorithms to solve it,
I design an algorithm for the problem and
I analyze the algorithm by proving it is correct and giving a
bound on the running time so as to establish the algorithms
efficiency.
Representative problems

I Interval scheduling: Given a set of jobs with start times and


finish times, find maximum cardinality subset of mutually
compatible jobs.
I Weighted interval scheduling: Given a set of jobs with start
times, finish times, and weights, find maximum weight subset of
mutually compatible jobs.
I Bipartite matching: Given a bipartite graph, find a maximum
cardinality matching.
I Independent set: Given a graph, find a maximum cardinality
independent set.
I A subset S V is independent if for every (u, v) E, either
u 6 S or v 6 S (or both).
Representative problems

I Interval scheduling: O(n log n) greedy algorithm.


I Weighted interval scheduling: O(n log n) dynamic programming
algorithm.
I Bipartite matching: O(nk ) network flow based algorithm.
I Independent set: NP-complete.

I There are problem even harder than NP-complete problems


Assignment

I We know that a stable matching is not always unique.


I Does G-S algorithm always return the same stable matching?
I Prove: Every execution of the G-S algorithm returns the same
stable matching S.
I Solved Exercises: 1, 2 (Chapter 1)
I Exercises: 3, 4, 5 (Chapter 1)

You might also like