Generating None-Plans in order to find plans

Abstract

We put forward a brand new approach to planning. The method aims at simplifying the task of planning in an abstract object oriented domain where entities are added only and never removed. Our approach is based on the synthesis of a family of all sets of actions that cannot be composed into a plan (called none-plans) in order to prune the state space searched for plans. We show how to build a propositional formula describing a set of the none-plans and how to approximate this set when the task of planning becomes too complex. A preliminary evaluation of the application of the none-plans synthesis to the generation of plans in the PlanICS framework is shown. The experimental results show a high potential of the approach.