FreeSML: delivering on the open-source simulation language promise
WSC '05 Proceedings of the 37th conference on Winter simulation
| Hi-index | 0.00 |
In computer-based discrete-event simulation (DES) implementations, a simulation executive manages the simulation clock and various collections of pending events. In addition to the future events list (FEL), the simulation executive may utilize a current events list (CEL) and one or more condition delay lists to manage simulated entities in different states. When all events enabled at the current simulated time have been processed, the simulation clock is advanced to the earliest scheduled time among the future events. When multiple events are enabled simultaneously, the simulation executive must select an order in which to execute them. Generally, their execution sequence is in descending order of their priority values, computed as of the time at which they are added to the pending list. One limitation of extant DES systems is that events in these pending lists cannot be re-ordered after they are scheduled. This is often an unrealistic constraint, because priorities for many real events vary as the system state evolves, and may include non-deterministic components. A novel methodology has been developed that removes this constraint and supports state-dependent and stochastic variation in the event execution sequence. This new methodology has been incorporated into a new discrete-event simulation software suite, FreeSML. Using this new tool, analysts can readily apply this new methodology to other models requiring state-dependent and stochastic event priorities.