CPEModeling and SimulationCPE 建模与仿真.ppt

CPE 345: Modeling and Simulation Lecture 3 Today’s topics Discrete event simulation General principles (chapter 3) Implementing a FIFO queue in OMNET++: introduction to OMNET++ programming Recall our simple bank teller example Concepts in Discrete-Event Simulation The two bank tellers example: revisited System state: Q(t) – number of customers waiting in the queue at time t S1(t) – state of teller 1: busy/idle S2(t) – state of teller 2: busy/idle Entities Teller 1 Teller 2 Arrivals generator (it is actually external to the system) Activities Teller 1 service time Teller 2 service time Inter-arrival time Delay Customer waiting time until service started or completed The two bank tellers example: events and discrete event simulation Events are determined by the completion of an activity – primary events: place event notice in FEL Service completion teller1 Service completion teller2 Inter-arrival time ends: new arrival occurs Delay completion – secondary events – not represented in FEL Other events: end of simulation Run for a preset time (e.g. 1000 seconds of simulator time) Run for a preset number of events (e.g. 1000 customers have been served) Run until an specific internal event occurs (e.g. until convergence of some simulated algorithm) Run until an external event occurs (e.g. stop by the user to modify experiment or to collect data) Discrete event simulation: produces a sequence of system snapshots (system images) which represent the evolution of the system through time. System snapshot example Event-scheduling/time advance algorithm FEL implementation Actions: Delete imminent event Removal of some events (cancellation for events) e.g. in wireless scenario, calls may be dropped, then, the end of service event for the dropped call must be cancelled Addition of new events e.g. after arrival of customer, new future arrival is scheduled Maintain time ordering of events If unordered list, a complete search for the imminent event must be done before each c