A Memory Adaptive Reasoning Technique for Solving the Audit Scheduling Problem

In this paper we propose a new heuristic search technique for solving the audit scheduling problem; it is based on memory adaptive reasoning. This technique, known as ART (Adaptive Reasoning Technique), augments the integer programming formulation of the audit scheduling problem with additional non-redundant constraints via memory-adjusting techniques to improve upon a starting feasible solution. The initial feasible solution is generated by a greedy forward loading procedure. The non-redundant constraints modify the execution of the greedy forward loading procedure by prohibiting poor choices that the greedy procedure makes.The performance of ART is compared with two well known heuristic procedures. The first is a recently developed tabu search (TS) procedure for solving the same problems and the other is a well known greedy randomized adaptive search procedure (GRASP). GRASP was modified to solve the audit scheduling problem. Application of the ART method to standard audit scheduling problems demonstrates a consistent improvement in the value of the cost itemization objective function compared to both TS and GRASP. In fact, it is shown that GRASP is a special case of ART and under certain conditions ART dominates TS