OR 335 - George Mason University

Download Report

Transcript OR 335 - George Mason University

OR 335 / SYST 335

Introduction to ARENA

Introduction

Brant Horio [email protected]

Back to the Basics

Fundamentals of simulation methodology are the cornerstone of all simulation packages Basics are necessary for accurate model development and model output understanding Different packages then become only an issue of a new syntax

Model Development is an Art!

Incredible model flexibility with ‘over the counter’ simulation software packages Art of practice involved with interpretation of system Endless possibilities in ‘how’ to model a specific system

Model Development is an Art!

Essential Core of Simulation Analysis is Scientific Experimentation

 Endless possibilities in ‘How’ to model a specific system  You must determine best way to account for all relevant assumptions/variables and maintain end-user flexibility  Maintain focus on problem as a whole

The Modeling Breakdown

10%:

Model Development

90%:

Problem Definition Data Collection / Analysis Scenario Experimentation Output Analysis Expected Success Measures Model Presentation

Success Story?

Aircraft Service Equipment Inventory Increase  Is the increase necessary?

 Is the expected service time improvement worth the associated massive capital expense?

These are questions Simulation Analysis is meant to answer

Success!

1 Simulation software license: Time to build model: Number of models built: Result: $18,000 3 months 1______ Priceless?

Nope, … Prevented capital expenditure of

~ $1 million

Introduction to ARENA

Windows driven interface Flowchart style model development  Process blocks  Process routing determined by connections between blocks  Templates contain process blocks Example model construction of a One-Node system

Example One-Node System

Model is composed of 3 parts    Creation/Introduction of entities to system Processing of entities in the system Exiting the entities from the system

Example Problem

The Sooty Dilemma

The Sooty Dilemma

CREATE:  43 new baby guinea pigs to be born at a rate of Triangular( 2, 5, 6 ) PROCESS:  Baby guinea pigs are adopted at a rate of Triangular( 3, 5, 8 ) Question:  What is the expected time for the entire litter to be adopted?

Note:

Distributions are unreasonable Assumptions of introduction of baby guinea pigs to system are inaccurate Emphasizes the point that:

GARBAGE IN, GARBAGE OUT!

Be Careful !!!

Basic Process

Template

Create

: Generate entities to system

Process

: Perform action on entities

Decide

: Routes entity based on condition

Dispose

: Exits entity from system

Additional Templates

CREATE

Module

Name

: Module name

Entity Type

: Unit to create for system

Time Between Arrivals

: Apply distribution or custom expression

Entities per Arrival

,

Max

and

First Creation

Spreadsheet Entry Available

Module parameter entry   Flowchart view ( Double-click module block ) Selection of module block also opens specific spreadsheet view on bottom of screen

Spreadsheet Modules

Spreadsheet displays based on template Create and manage model components in centralized area

When in Doubt, Right Click

‘Build Expression’ provides flexibility Most useful for applying distributions Generate random numbers Create custom formulas

PROCESS

Module

Action:

   

Delay:

dd

Seize Delay:

dd

Seize Delay Release:

dd

Delay Release:

sds

PROCESS Module ‘Action’ Options

Delay

: Delays/Processes entity for a given time period ( No Queue )

Seize Delay ( Part 1 )

: Utilizes resource to process entity ( Queue )

Delay Release ( Part 2 )

: Utilizes captured resource to process, then releases resource

Seize Delay Release

: Utilizes resource to process entity, then allows resource to process another entity ( Queue )

PROCESS Module - Queues

Queues are customizable ( Dbl Click Queue )

DISPOSE

Module

Exits entity from system Automatically tracks entity throughput

Run Setup

Model Run

Remember time units!!!

General rule of thumb for replication is 30 Model will have automatic basic animations After run completion; report generation

Output Reports

Category by Replication  Most data output is automatic Litter to be completely adopted in 272 days based on current assumptions!

ANIMATION!

Focus on basics for now Power of simulation is best embodied by its ability to communicate and justify recommendations to decision makers ‘Buy Off’ is of critical importance Animation is invaluable in assisting communication of the model May also act as a debugging tool

ANIMATION

Entity animation pictures are customizable under EDIT / ENTITY PICTURES Resource animation pictures are customizable through the ‘Animate’ button

Happy Modeling

Play around with Arena on your own Increased familiarity = increased model development proficiency

Questions?