A New Algorithm for Planning with Resource-incompatible Actions

A new approach to handle resource-constraint planning problems with resource-incompatible actions is proposed in this paper. Some new concepts are defined, such as critical resource variable, resource variable flag and resource correlative actions etc. A new action effect called temporal effects is studied. Based on these new definitions, a novel algorithm is provided for resource-constraint planning problems with resource-incompatible actions. This algorithm can deal with more complex resource-constrained planning problems

I. INTRODUCTION

Intelligent planning has been a pop domain in Artificial Intelligent domains, especially after the birth of Graphplan ^[1]. A lot of studies about Graphplan have appeared recently, such as literature [2]–[4]. After the extension of the Graphplan formalism to ADL, another important aspect of planning problems forms the focus for the development of IPP ^[5]: resource-constrained planning problems. These problems lead resource effects and resource requirements to classical planning problems. Jana Koehler provides formal definition of resource effects and resource requirements in literature [6], and presents a searching algorithm based on Graphplan. Consequently a method solving resource-constrained problems is proposed. Jana Koehler's method can solve problems with resource-compatible actions effectively. There are a lot of resource-incompatible planning problems in real world. For planning with resource-incompatible actions, the method in literature [6] can not find valid plan.