Automated constraint-based scheduling for pilot training

The author discusses the functional problems characteristic of automated scheduling for the integrated training of military pilots, including the size of the search space (which is combinatorially explosive due to complexity and number of constraints), multiple competing goals, the time-based nature of the problem, and multiagent elements. She compares optimal linear programming approaches to the problem with artificial intelligence techniques. The solution techniques are explained, the advantages of an expert system approach are examined, and the approach, design, and performance of the system are discussed