Disassembly process planning using Petri nets

We generate a disassembly PN (DPN) from a disassembly precedence matrix. The resulting DPN can be analyzed using the reachability tree method to generate all feasible disassembly process plans (DPPs), and cost functions can be used to determine the optimal DPP. Since generating the reachability tree is NP-complete, we develop a heuristic algorithm to limit the size of the reachability tree. The algorithm employs multi-hypothesis search to dynamically explore the /spl nu/ likeliest lowest cost branches of the tree, in order to identify near-optimal DPPs. The cost function incorporates tool changes, changes in direction movement, and individual part characteristics (e.g., hazardous). An example is used to illustrate the procedure.