Parallel Proposals in Asynchronous Search

Parallelism and distribution are two distinct concepts that are confusingly close. Parallel Search refers in this work to the distribution of the search space and Distributed Asynchronous Search to the distribution of the constraint predicates. A certain amount of parallelism exists in any Distributed Asynchronous Search and it increases with the degree of asynchronism. However, in comparison to Parallel Search [10], the parallel effort in Distributed Asynchronous Search can be more redundant. Moreover, agents in Asynchronous Search can have periods of inactivity which are less frequent in Parallel Search. Since Distributed Search is the only solution for certain classes of naturally distributed problems, we show here how one can integrate the idea of Parallel Search in Distributed Asynchronous Search. A technique for dynamic reallocation of search space is then presented. This technique builds on the procedure for marking concurrent proposals for conflicting resources, that we have formalized in [11].