IE360 Industrial System

Simulation (Th)
Lecture Set 01 (Duration 2hrs)

Prepared by
Dr. Shahid Maqsood

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 1
 Contact Details
 Prof. Dr. Shahid Maqsood
 Email: [email protected] and
[email protected]
 Ph: 03338900098
 (Only Text with your name at bottom of text will be answered)

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 2
 Grading

 Final Exam: 50%

 Midterm : 25%
 Assign/Quiz/Attend/project etc. 25%

 Facebook Official Group (For Updates/Announcements)

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 3
Bonus Points

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 4
 Course Outline
 Concepts of discrete-event system modelling and simulation.
 Use of the commercial Arena software.
 Statistical models in simulations.
 Material flow models: Assembly lines, transfer lines, Flexible
manufacturing systems, shop scheduling with many products.
 Inventory control and Just-in-Time simulations.
 Analysis of simulation data : Input modelling, Verification and
validation of simulation models, output modelling and analysis,
comparison and evaluations of alternative system designs.
 Random Numbers: random number generations, random
variate generation.
© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/5
 Lecture 1

Introduction to Industrial Systems Simulation

- What is Simulation -

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/6
 Simulation Is ...
 Very broad term, set of problems/approaches
 Generally, imitation of a system via computer
 Involves a model—validity?
 Don’t even aspire (ultimate goal or desire) to analytic solution
 Don’t get exact results (bad)
 Allows for complex, realistic models (good)
 Approximate answer to exact problem is better than exact
answer to approximate problem
 Consistently ranked as most useful, powerful of mathematical-
modeling approaches

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Validation
 In computer modelling and simulation, the process of
determining the degree to which a model or
simulation is an accurate representation of the real
world from the perspective of the intended uses of the
model or simulation.

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 8
 Some Application Areas

 Manufacturing—scheduling, inventory
 Staffing personal-service operations
 Banks, fast food, theme parks, Post Office, ...
 Distribution and logistics
 Health care—emergency, operating rooms
 Computer systems
 Telecommunications
 Military
 Public policy
 Emergency planning
 Courts, prisons, probation/parole

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Systems

 Physical facility/process, actual or planned

 Study its performance
 Measure
 Improve
 Design (if it doesn’t exist)
 Maybe control in real time
 Sometimes possible to “play” with the system
 But sometimes impossible to do so
 Doesn’t exist
 Disruptive, expensive

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Models

 Abstraction/simplification of the system used as a proxy for the

system itself
 Can try wide-ranging ideas in the model
 Make your mistakes on the computer where they don’t count, rather
for real where they do count
 Issue of model validity
 Two types of models
 Physical (iconic) (x)
 Logical/Mathematical—quantitative and logical assumptions,
approximations

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 What Do You Do with a Logical Model?

 If model is simple enough, use traditional mathematics

(queuing theory, differential equations, linear
programming) to get “answers”
 Nice in the sense that you get “exact” answers to the model
 But might involve many simplifying assumptions to make the model
analytically tractable—validity??
 Many complex systems require complex models for
validity—simulation needed

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Computer Simulation
 Methods for studying a wide variety of models of real-
world systems
 Use numerical evaluation on computer
 Use software to imitate the system’s operations and characteristics,
often over time
 In practice, is the process of designing and creating
computerized model of system and doing numerical
computer-based experiments
 Real power—application to complex systems
 Simulation can tolerate complex models

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Popularity
 M.S. grads, CWRU O.R. Department (1978)
 Asked about value after graduation; rankings:
1. Statistical analysis, 2. Forecasting, 3. Systems analysis, 4. Information
systems
5. Simulation
 137 large firms (1979)
1. Statistical analysis (93% used it)
2. Simulation (84%)
 Followed by LP, PERT/CPM, inventory, NLP

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Popularity (cont’d.)

 (A)IIE, O.R. division members (1980)

 First in utility and interest: Simulation
 But first in familiarity: LP (simulation was second)
 Longitudinal study of corporate practice (1983, 1989,
1993)
1. Statistical analysis
2. Simulation
 Survey of such surveys (1989)
 Consistent heavy use of simulation

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/16
 Advantages of Simulation

 Flexibility to model things as they are (even if messy

and complicated)
 Avoid “looking where the light is” (a morality play):
You’re walking along in the dark and see someone on hands and knees searching the ground under a
street light.
You: “What’s wrong? Can I help you?”
Other person: “I dropped my car keys and can’t find them.”
You: “Oh, so you dropped them around here, huh?”
Other person: “No, I dropped them over there.” (Points into the darkness.)
You: “Then why are you looking here?”
Other person: “Because this is where the light is.”

 Allows uncertainty, non-stationarity in modeling

 The only thing that’s for sure: nothing is for sure
 Danger of ignoring system variability
 Model validity
© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Advantages of Simulation (cont’d.)

 Advances in computing/cost ratios

 Estimated that 75% of computing power is used for various kinds of
simulations
 Dedicated machines (e.g., real-time shop-floor control)
 Advances in simulation software
 Far easier to use (GUIs)
 No longer as restrictive in modeling constructs (hierarchical, down to
C)
 Statistical design & analysis capabilities

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 The Bad News

 Don’t get exact answers, only approximations, estimates

 Also true of many other modern methods
 Can bound errors by machine roundoff
 Get random output (RIRO) from stochastic simulations
 Statistical design, analysis of simulation experiments
 Exploit: noise control, replicability, sequential sampling, variance-
reduction techniques
 Catch: “standard” statistical methods seldom work

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 The Bad News (cont’d.)

 Cannot optimize any system’s performance.

 Cannot give accurate results if the model and
input data are inaccurate.
 Cannot describe system characteristics that have
not been explicitly modelled.
 Cannot Solve any problems (can only provide
information, from which solutions can be
inferred).
© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Different Kinds of Simulation

 Static vs. Dynamic

 Does time have a role in the model?
 Continuous-change vs. Discrete-change
 Can the “state” change continuously or only at discrete points in time?
 Deterministic vs. Stochastic
 Is everything for sure or is there uncertainty?
 Most operational models:
 Dynamic, Discrete-change, Stochastic

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 In Manufacturing Systems

 Purely Discrete
 Parts Arriving and leaving at specific times
 Machine going down and up at specific times
 Breaks for Workers
 Mixed Continuous-discrete models
 Refinery with continuous changing pressure inside vessels and
discretely occurring shutdowns

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 22
SIMULATION BY HAND

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 23
 The Buffon Needle Problem

 Estimate p (George Louis Leclerc, c. 1733)

 Toss needle of length l onto table with stripes d (>l)
apart
2l
 P (needle crosses a line) =
pd
 Repeat; tally p = proportion of times a line is crossed
 Estimate p by 2l

pd

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/24
 Why Toss Needles?
 Buffon needle problem seems silly now, but it has
important simulation features:
 Experiment to estimate something hard to compute exactly (in 1733)
 Randomness, so estimate will not be exact; estimate the error in the
estimate
 Replication (the more the better) to reduce error
 Sequential sampling to control error—keep tossing until probable error in
estimate is “small enough”
 Variance reduction (Buffon Cross)

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Using Computers to Simulate
 General-purpose languages (FORTRAN)
 Tedious, low-level, error-prone
 But, almost complete flexibility
 Support packages
 Subroutines for list processing, bookkeeping, time advance
 Widely distributed, widely modified
 Spreadsheets
 Usually static models
 Financial scenarios, distribution sampling, SQC

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/26
 Using Computers to Simulate (cont’d.)

 Simulation languages
 GPSS, SIMSCRIPT, SLAM, SIMAN
 Popular, in wide use today
 Learning curve for features, effective use, syntax
 High-level simulators
 Very easy, graphical interface
 Domain-restricted (manufacturing, communications)
 Limited flexibility—model validity?

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/27
 Where Arena Fits In
 Hierarchical structure Higher User-Created Templates
Commonly used constructs

 Multiple levels of modeling Company-specific processes

Company-specific templates
A single
etc.
 Can mix different modeling graphical user
interface

levels together in the same

Vertical Solutions
consistent at
Application Solution Templates any level of

model Call$im
BP$im
etc.
modeling

 Often, start high then go lower

as needed Common Panel

Professional Edition
Many common modeling constructs

Get ease-of-use advantage of

Very accessible, easy to use

Arena Template
 Level of
Reasonable flexibility

simulators without sacrificing

Modeling

Support, Transfer Panels

modeling flexibility Access to more detailed modeling for greater

Standard Edition
flexibility

SIMAN Template
Blocks, Elements Panels
All the flexibility of the SIMAN simulation
language

User-Written Visual Basic, C/C++, FORTRAN

Code
The ultimate in flexibility
Lower C/C++/FORTRAN requires compiler

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 When Simulations are Used
 Uses of simulation have evolved with hardware,
software
 The early years (1950s-1960s)
 Very expensive, specialized tool to use
 Required big computers, special training
 Mostly in FORTRAN (or even Assembler)
 Processing cost as high as $1000/hour for a sub-286 level machine

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/29
 When Simulations are Used (cont’d.)

 The formative years (1970s-early 1980s)

 Computers got faster, cheaper
 Value of simulation more widely recognized
 Simulation software improved, but they were still languages to be
learned, typed, batch processed
 Often used to clean up “disasters” in auto, aerospace industries
 Car plant; heavy demand for certain model
 Line underperforming
 Simulated, problem identified
 But demand had dried up—simulation was too late

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/30
 When Simulations are Used (cont’d.)

 The recent past (late 1980s)

 Microcomputer power
 Software expanded into GUIs, animation
 Wider acceptance across more areas
 Traditional manufacturing applications
 Services
 Health care
 “Business processes”
 Still mostly in large firms
 Often a simulation is part of the “specs”

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan lect1/31
 When Simulations are Used (cont’d.)

 The present
 Proliferating into smaller firms
 Becoming a standard tool
 Being used earlier in design phase
 Real-time control
 The future
 Exploiting interoperability of operating systems
 Specialized “templates” for industries, firms
 Automated statistical design, analysis

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan
 Concluding Remarks
 For Our Money
 We Would prefer an approximate answer to the right problem rather
than an exact answer to the wrong problem
 This is Called TYPE III Error
 Providing the right answer to the wrong question

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 33
That’s All for Today

© Dr. Shahid Maqsood, Assistant Professor, Department of Industrial Engineering, UET Peshawar, Pakistan 34