We introduce a new ontology, a modeling approach to planning in STRIPS-like domains, and an extended numeric framework based on converting planning problems into Mixed-Integer Programming problems (MIPs). We describe a new planner, OptiPlan, which is based on that ontology. Frugal by design, OptiPlan always finds the provable "cheapest" plan or proves that the problem is infeasible. Extensions of STRIPS-like domains are applicable to a wide variety of Supply Chain Problems and thus build a bridge between Artificial Intelligence Planning and Manufacturing Planning. Furthermore, the new ontology offers an efficient, constructive alternative to a fifty-year old Bill-of-Materials (BOM) and Routing modeling philosophy with the controversial issue of BOM/Routing separation.
展开▼