Analogy-based multiple process planning system with resource conflicts

Computer-aided process planning is becoming a widely prevalent technology in modern manufacturing systems. The research presented here describes a new methodology for generating process plans based on the analogy deductive paradigm. The method uses rules that represent relations between two shapes and allow inference of the type: shape A is to shape B as C is to D, where usually D is the unknown shape. The system uses backward chaining and therefore gradually converts the part from its finished (designed) form into its initial form. This method can generate multiple process plans for each given part; the paper also presents a method of selecting the best combination of process plans to maximize the production rate of that part. Once the dominant combination of plans is selected, the paper presents a method to calculate a proper production quantity for each process plan. This method is based on `coalition theory' and uses Shapley values to evaluate each member of such a coalition. The system has been implemented on a SUN workstation using Quintus Prolog and C++. The current implementation considers prismatic parts only.