Offline and Online Plan Library Maintenance in Case-Based Planning

One of the ways to address the high computational complexity of planning is reusing previously found solutions whenever they are applicable to the new problem with limited adaptation. To do so, a reuse planning system needs to store found solutions in a library of plans, also called a case base. The quality of such a library critically influences the performance of the planner, and therefore it needs to be carefully designed and created. For this reason, it may be also important to update the library during the lifetime of the system, as the type of problems being addressed may evolve or differ from the ones the case base was originally designed for. In our ongoing research, we address the problem of maintaining the library of plans in a recent case-based planner called OAKplan. After having developed offline techniques to reduce an oversized library, we introduce here a complementary online approach that attempts to limit the growth of the library, and we consider the combination of offline and online techniques to ensure the best performance of the case-based planner. The different investigated approaches and techniques are then experimentally evaluated and compared.