Automatically Selecting and Using Primary Eeects in Planning: Theory and Experiments

The use of primary eeects of operators is an eeective approach to improving the eeciency of planning. The characterization of \good" primary eeects, however, has remained at an informal level and there have been no algorithms for selecting primary eeects of operators. We formalize the use of primary eeects in planning and present a criterion for selecting useful primary eeects, which guarantees eeciency and completeness. We analyze the eeciency of planning with primary eeects and the quality of the resulting plans. We then describe a learning algorithm that automatically selects primary eeects and demonstrate, both analytically and empirically, that the use of this algorithm signiicantly reduces planning time and does not compromise completeness.