Chapter 1

What is Simulation? 

What is Simulation?

Simulation refers to a broad collection of methods and applications to mimic the behavior of real systems, usually on a computer with appropriate software.

1.1 Modeling

Modeling a process facilitates analysis. Thus, to make improvements to a process, we need a model of that process.

1.1.1 What’s Being Modeled?

We model processes.

What are examples of processes we model?

See page 4

Why do we model?

1. To improve performance of a process
2. To design a process if it does not exist
3. To manage a process on a daily basis
4. To understand a current process and its performance

• To build teamwork
• To create buy-in to the new process and to simulation

1.1.2 How About Just Playing With The System?

We could just make changes to the actual process and not do a simulation.

When is "Just Playing With The System" appropriate?

1. When the solution is obvious (a direct cause and effect relationship - we know with great certainty that a change will produce an outcome)
2. When the cost of conducting the experiment is low and the consequences of failure are low
3. When the experiment is "controlled" - variables are carefully measured and monitored

What are the advantages of "just playing with the process"?

1. Cheap compared to a full simulation
2. Quick compared to a full simulation

What are the disadvantages of "just playing with the process"?

1. Risk costly mistakes - an error causes harm to the process such as more defects, longer cycle time, etc..
2. Overlook the best solution - because we did not try every alternative, we just tried one or two

1.1.3 Sometimes You Can’t (or Shouldn’t) Play With the Process

1. Cost of failure is too high
2. If changes are large, then the cost of making changes could be high

What are the advantages of building a model?

1. We can do what "if analysis"
2. Low risk in making changes
3. Better chance of implementing the "best" solution (most optimal)
4. Facilitates process design
5. Requires data collection, which brings understanding of the process

We must develop models with enough detail so that what you learn about the process will be reflected in the real process when changes are implemented.

1.1.4 Physical Models

Iconic model - a physical replica or scale model of the system

What are examples of physical models?

1. Tabletop models
2. Full scale models
3. Physical flight simulators

1.1.5 Logical (or mathematical models)

What is a logical model?

A set of approximations and assumptions, both structural and quantitative, about the way the system does or will operate

Mathematical models are usually represented on a computer.

What are the advantages of a logical model?

1. Easy
2. Cheap
3. Fast
4. Make mistakes on the computer where the consequences of failure are low

1.1.6 What Do You Do with a Logical Model?

If the process is simple, we can use queuing theory, differential equation methods or linear programming.

If the process is complicated, then we must use simulation.

1.2 Computer Simulation

What is computer simulation?

The process of designing an creating a computerized model of a real or proposed system for the purpose of conducting numerical experiments to provide a better understanding of the behavior of that process or a given set of conditions.

1.2.1 Popularity and Advantages

What are the advantages of computer simulation?

1. Can model complicated processes
2. Improvement in price and performance of computer hardware
3. Improvement in simulation software - power, flexibility, ease of use - no more tedious and error pone programming

1.2.2 The Bad News

Real systems have uncontrollable and random inputs, which causes the outputs to be random too. We see something different every time.

To over come this problem, we run the simulation many times to see what happens over time. The question is how many times do we run the simulation?

We could make a lot of over-simplifying assumptions, which would produce non-random results. This approach may lead to an over simplified solution.

We should strive for an approximate solution to the right problem rather than the exact solution to the wrong problem.

1.2.3 Different Kinds of Simulations

1. Static vs. dynamic –

• static – no time element
• dynamic – models process over time

2. Continuous vs. discrete

• Continuous – state o he model changes continuously over time such as the level of a reservior as water flows in and out
• Discrete – change can occur ay specific points in time such as parts arriving and departing in a manufacturing plant

3. Deterministic vs. Stochastic

• Deterministic – models that have no random input such as appointment book operations
• Stochastic – operate with random input such as people entering a bank

Models can have both deterministic and stochastic aspects

1.3 Simulation Gets Done

1.3.1 By Hand

The first attempts at simulation were by hand. They used the tools they had at the time.

1.3.2 Programming In General-Purpose Language

Building simulations using a generic programming language such as FORTRAN, C, C++ or others.

What are the advantages?

1. Highly customizable
2. Very flexible

What are the disadvantages?

1. Tedious
2. Error Prone
3. Time consuming
4. Requires skilled programmers

1.3.3 Simulation Languages

Computer languages designed specifically or simulation.

Examples:

• GPSS
• SIMSCRIPT
• SLAM
• SIMAN

What are the advantages?

1. Faster than general purpose languages
2. Includes many features that reduce the time required to develop a model

What are the disadvantages?

1. Time consuming
2. Requires skilled programmers
3. Cryptic users interfaces and syntax

1.3.4 High-Level Simulators

Windows operated software packages that have mouse driven user interfaces, menus and dialogs.

Users select from construct, connect them and run the model with a dynamic graphical animation of process components.

What are the advantages?

1. Easy to use
2. Fast model development
3. Users have short learning curve - thus many people can learn to use it
4. Dynamic graphical animation helps communicate ideas

What are the disadvantages?

1. Limited flexibility - (though enough flexibility to handle most situations)

Notice how short this list is!!!

1.3.5 Where Does Arena Fit In?

Arena has templates of modules that allow the user to increase flexibility when needed.

1.4 When Simulations Are Used

1.4.1 The Early Years (1950-1969)

Simulation was very expensive and specialized

1.4.2 The Formative Years (1970 - 1989)

More companies began to use simulation because computer hardware and software became more affordable.

1.4.3 The Recent Past (1980’s)

Simulation was being used to justify capitol investments

1.4.4 The Present (1990 - present)

1. Better animation
2. Ease to use
3. Faster computers
4. Easy integration with other packages
5. Being used earlier in the design phase
6. Used to monitor process performance

1.4.5 The Future

1. More integration with other packages
2. More vertical applications (more specific)
3. Models will become living models