Concurrent Search

Abstract

As noted in Chapter 1, search had a prominent role in the beginning of artificial intelligence. Early textbooks on the subject (e.g. [Nilsson, 1971]) devote a great deal of space to search. Improved search methods were considered worthy subjects for inclusion in artificial intelligence conferences and journals. While consideration of search in the abstract has tended to move from the field of artificial intelligence to mainstream computer science and operational research, search continues to be an important tool for artificial intelligence practitioners. Parallel and distributed search continues to be an active field of research in computer science, much of the interest coming from the fact that search algorithms designed for sequential computers rarely map easily onto distributed and parallel architectures. The challenge for programming language design is to invent languages which free the programmer from having to consider low level details of the parallel architecture. This allows the programmer to concentrate on the more abstract aspects of concurrent algorithms, while retaining sufficient control to efficiently exploit the parallelism available in the architecture.