Knowledge-based allocation model for distributed processing systems

The authors propose a model for a knowledge-based assistant (expert system) to aid in allocating a blackboard-oriented system to a multiprocessor or distributed platform. It incorporates variants of several important techniques of artificial intelligence to provide recommendations for the allocation of a set of knowledge sources to a set of processors. To provide intelligent recommendations, an expert system acquires the same information about the two domains as an analytical allocation algorithm. However, while most allocation algorithms are limited to finding a minimum cost solution, the expert scheduler considers the complexities of constraints and multiple objectives of the practical system and/or application.<>