Planning with Disjunctive Temporal Constraints

Many real-world problems contain elements of both planning and scheduling problems. In response, many frameworks and systems have been developed that add scheduling abilities to planning systems. We continue this trend by presenting a framework and algorithm for planning with disjunctive temporal constraints. We define a constraint-based interval (CBI) framework for describing a planning problem with disjunctive temporal constraints, which we call a disjunctive temporal planning problem (DTPP). We also describe the Disjunctive Temporal Partial-Order Planner (DT-POP), a planning algorithm that is capable of solving DTPP’s, and analyze several techniques for improving its efficiency. Although the current implementation of DT-POP is not yet competitive with other planning systems in terms of efficiency, it is more expressive in terms of complex temporal constraints. We offer several ideas for improving the efficiency of DTPOP in the future, in hopes that the current work will be the beginning of an interesting line of research.