ON-LINE RESCHEDULING OF RAPID TRANSIT SYSTEMS : A FUZZY EXPERT SYSTEM APPROACH

Abstract

The performance of a rapid transit system, in terms of passenger service and operation cost, depends crucially on the traffic control system. Without any form of traffic control, the performance of a rapid transit system which functions according to pre-determined time-tables is easily affected by operational disturbances. Conversely, if a responsive traffic control system is implemented, the effect of these disturbances can be avoided or reduced. Traffic control can be achieved by rescheduling train services. The purpose of this research project is to devise an on-line rescheduling strategy that optimizes passenger service at low operation cost. The on-line rescheduling strategy comprises of two rescheduling algorithms. One of the algorithms activates the command that reschedules the despatch of trains from terminals, whilst the other algorithm activates the command that reschedules train departure from passenger stations. The concept of a fuzzy expert system is used in these algorithms due to its ability in handling complex decision making under indeterminate conditions. Object-oriented techniques are used to simulate the operation of the rapid transit system. The techniques allow each object in the simulation model to have the attributes, operations, and structures of its counterpart in the real world. This feature resolves the complexity of knowledge representation in the rapid transit system. To evaluate the performance of the on-line rescheduling strategy, the on-line rescheduling algorithms are also represented by software objects embedded in the software module representing the existing Automatic Train Supervision System (ATS). The simulation tests reveal that the on-line rescheduling strategy is capable of optimizing the performance of the rapid transit system when disturbances occur.