TRAPS - A Time Dependent Resource Allocation Language

A general language for specifying resource allocation in Scheduling problems is presented. TRAPS is a superclass of the RAPS (Resource Allocation Problem Specication) language developed by the authors [31]. TRAPS enables the speci cation of a scheduling problem by adding built in time operators, on top of existing terms for resources, activities, allocation rules and constraints. In this way TRAPS provides a convenient knowledge acquisition tool. The language syntax is powerful and allows the speci cation of rules and constraints which are di cult to formulate with traditional approaches, and it also supports the speci cation of various control and backtracking strategies. The generalized inference engine that runs compiled TRAPS programs is enhanced to provide all needed operations for a typical PERT/CPM calculations on a schedule. This engine acts as an expert system shell and is called ESRA (Expert System for Resource Allocation). The performance of TRAPS combined with ESRA is demonstrated by analyzing its solution of a typical scheduling problem, that of a software management project. The analysis shows that certain heuristics which formed the basis of our strategy for resource allocation (cf. [31]) perform very well also in the (time dependent) scheduling domain. keywords: expert systems, scheduling, resource allocation, Prolog, knowledge acquisition.