An experiment in scheduling and planning of non-structured jobs: Lessons learned from artificial intelligence and operational research toolbox

Most scheduling problems traditionally address well defined and structured environments. Some examples include manufacturing (job shop, flow shop, etc.) and project scheduling respectively. Another type of scheduling problem that has received little or no attention is defined here as a non-structured scheduling problem (NSSP). A typical NSSP addressed here involves scheduling aircraft turnaround functions. The scheduling method consists of artificial intelligence and operational research techniques. The results obtained from the hybrid model indicate that flexibility and knowledge replication can be achieved at various levels of abstraction by converting non-structured problems to their structured equivalents. The model is implemented with a Task Oriented Planner, a decision support system for multiagent task scheduling and planning.