Simulation - Palgrave仿真与.pptVIP

Chapter 17 Simulation Types of simulation Discrete event (or queuing) simulation Continuous simulation Monte Carlo (or spreadsheet) simulation Discrete event simulation If the system to be modelled can be represented by a series of discrete events, the method known as discrete event simulation can be used. Systems that can be modelled using the discrete event approach are generally systems where queuing mechanisms operate. Continuous simulation A system is modelled using differential equations. These equations usually need to be solved numerically. Continuous simulation is normally applied where the approximation of events occurring at discrete intervals of time is not valid (e.g. fluid flow). Monte Carlo simulation Monte Carlo simulation is often reserved for simulation models where queuing is not an integral part of the system. Examples include inventory problems and risk analysis. These types of systems can usually be modelled using a spreadsheet. The modelling cycle Formulate the problem and collect the data Build the model of the system Verify (or debug) the computer model Validate the model Make runs of the model to answer 憌hat-if?types of questions Random numbers Most simulation models are stochastic in nature and to replicate the random nature of these systems random numbers are used. Random numbers generated by a computer are called pseudo random numbers. Random numbers are used to generate probability distributions, which define how the inputs to the model vary. Types of systems Terminating systems A terminating system is one where the system ends or terminates after a period of time of some other condition. Non-terminating systems A non-terminating system is a system where the system does not terminate or where the period is so long that it is effectively non-terminating. Validation of simulation models The basic principle of validation is that the model should replicate the real system reasonably closely. Output from t