The W-6 approach to multi-dimensional scheduling: where AI and operations research meet

Inherent limitations of the traditional approaches to scheduling and allocation of resources are pinpointed, and new directions are proposed. These are based on five years experience with very-large-scale scheduling problems that resulted in a working software system called W-6. W-6, which represents a compromised approach that may not generate a globally optimal solution but has proven to lead to practical solutions, is described. The key elements of the W-6 approach are: shifting the emphasis from search space exploration to problem decomposition, and a natural language interface that allows the user to formulate the problem requirements, constraints, and goals in a reasonably restricted natural language.<
>