Highly parallelizable route planner based on cellular automata algorithms by P. N. Stiles I. S. Glickstein An overview Is presented of our work on a highly parallelizable route planner that efficiently finds an optimal route between two points; both serial and massively parallel Implementations are described. We compare the advantages and disadvantages of the associated search algorithm relative to other search algorithms, and conclude with a discussion of future extensions and related applications. Introduction Route planning addresses the question of how "best" to go from point A to point B. Solutions are important for routing and scheduling manned and unmanned air, land, and sea vehicles. The best route may be the one that is fastest, safest, cheapest, or smoothest, or it may be the one that covers the largest search area. It usually is one that optimizes some combination of these factors. The famous traveling salesman problem is NP-complete; i.e.. there is no known polynomial time algorithm for inding an optimal solution. The shortest-path route-planning problem has polynomial complexity which is not as bad, but still poses a challenge for today's computers operating on real- time applications. This overview covers various aspects of the route-planning problem and focuses on algorithms we have developed and associated insights we have gained over the past eight years. Route-planning example Consider the pilot of a medical rescue helicopter on a stormy night who needs to plan a route from a hospital to the scene of an accident. The pilot wants to avoid the worst areas of the storm but also must arrive quickly to save the patient. There is likely to be a conflict between a safer route that avoids the worst of the storm but takes longer, and a short, direct route that is faster but more dangerous. The pilot also wants to avoidflyingnear radio towers, transmission lines, and developed areas with tall buildings, since the reduced visibility makes them •Cow^Wit 1994 by International Business Machines Corporation. Copying in printed fonn for private use is pennilted without payment o( royalty provided that (I) each reproduction is done widioat alteration and (2) SheJcmmal reference and IBM copyright notice are inctoded on the first page. The title and abstract, but no other portions, of this paper may be copied or distributed royalty ftee without tather permission by computer-based and other inforniation.service systems. Permission to republish any other portion of this paper must be obtained from the Editor. 167 IBM J. RES. DEVEUaP. VOL. 38 NO. 2 MARCH 1994 P. N. STILES AND I. S. GLICKSTEIN