Solving the weighted region least cost path problem using transputers.

Abstract

The weighted region least cost path problem involves finding the minimal cost path between a source point and a goal point through a plane that has been subdivided into weighted regions. In this thesis, we investigate a new parallel approach which seeks to take advantage of the distributed, asynchronous computing environment provided by the INMOS Transputer. The algorithm consists of a family of local, asynchronous, iterative and parallel procedures. The program is implemented on a network of transputers using a parallel version of the C programming language and tested on various maps of triangulated weighted regions. Results were favorable in terms of producing a near optimum path and reduced processing time.