Understanding Simulation

Computer Simulation Uses

General Situation Real system does not yet exist and building a prototype is cost-prohibitive, time-consuming or hazardous System is impossible to build Real system exists but experimentation is too expensive, hazardous or disruptive to conduct Forecasting is required to analyze long time periods in a compressed format Mathematical modeling has no practical analytical or numeric solution Examples Aircraft, Production System, Nuclear Reactor National Economy, Biological Systems Proposed changes to a Materials Handling System, Military Unit, Transportation System, Airport Baggage Handling System Population Growth, Forest Fire Spread, Urbanization Studies, Pandemic Flu Spread Stochastic Problems, Nonlinear Differential Equations

General Benefits of Simulation

Allows Experimentation without Disruptions to Existing Systems Concept can be Tested Prior to Installation Detection of Unforeseen Problems or Bugs Speed in Analysis Forces System Definition Enhances Creativity

General Limitations of Computer Simulation

Expensive Time-Consuming Yields Only Approximate Answers Difficult to Validate Accepted as Gospel

Different Types of Simulation

Continuous Simulation Monte Carlo Simulation Discrete Event Computer Simulation Agent-Based Modeling

Continuous Simulation
May consist of algebraic systems, game theoretic models, statistical models or differential equations. Represents the recession and flow of parameters associated with the system state. Example : Predator-Prey Model

Monte Carlo Simulation

Invokes the thought of gambling, gaming and chance. Represent simulations that are a scheme employing random numbers, which is used for solving stochastic or deterministic problems where the passage of time plays no role.

Discrete Event Simulation

Characterized by the passage of blocks of time during which nothing happens, punctuated by events which chance the state of the system. Example: ATM Simple queuing system (determine maximum queue length, average service wait time, and percent time the ATM is in use)

Agent-Based Modeling
Addresses the simultaneous interactions of multiple agents to simulate, recreate, study, and predict complex phenomenon. Over-all behavior emerges through micro-level interactions of individual agents.

ABM: Elements
Multiple agents modeled and scaled with various levels of detail (granularity) Decision-making heuristic rules Adaptive behaviors or learning Interaction rules or topology Environment for interaction often consisting of constrained resources.

Simulation Languages Features

Simulation clock or mechanism for advancing simulated time. Methods to schedule the occurrence of events Tools to collect and analyze statistics concerning the usage of various resources and entities. Tools for reporting results Debugging and error detection facilities Random numbers generators and related sets of tools General frameworks for model creation