Integrating planning and learning: the PRODIGY architecture

Abstract Planning is a complex reasoning task that is well suited for the study of improving performance and knowledge by learning, i.e. by accumulation and interpretation of planning experience. PRODIGY is an architecture that integrates planning with multiple learning mechanisms. Learning occurs at the planner's decision points and integration in PRODIGY is achieved via mutually interpretable knowledge structures. This article describes the PRODIGY planner, briefly reports on several learning modules developed earlier along the project, and presents in more detail two recently explored methods to learn to generate plans of better quality. We introduce the techniques, illustrate them with comprehensive examples, and show preliminary empirical results. The article also includes a retrospective discussion of the characteristics of the overall PRODIGY architecture and discusses their evolution within the goal of the project of building a large and robust integrated planning and learning system.

[1]  G. Pólya,et al.  How to Solve It , 1945 .

[2]  Lawrence E. Doyle,et al.  Manufacturing processes and materials for engineers , 1969 .

[3]  Richard Fikes,et al.  STRIPS: A New Approach to the Application of Theorem Proving to Problem Solving , 1971, IJCAI.

[4]  David Warren,et al.  Warplan: a system for generating plans , 1974 .

[5]  Nils J. Nilsson,et al.  Artificial Intelligence , 1974, IFIP Congress.

[6]  Austin Tate,et al.  Generating Project Networks , 1977, IJCAI.

[7]  Pat Langley,et al.  Learning Effective Search Heuristics , 1983, IJCAI.

[8]  J. Ross Quinlan,et al.  Learning Efficient Classification Procedures and Their Application to Chess End Games , 1983 .

[9]  Allen Newell,et al.  R1-Soar: An Experiment in Knowledge-Intensive Programming in a Problem-Solving Architecture , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[10]  Jean-Claude Latombe,et al.  Making Compromises Among Antagonist Constraints in a Planner , 1985, Artif. Intell..

[11]  Dana S. Nau,et al.  Hierarchical representation of problem‐solving knowledge in a frame‐based process planning system , 1986, Int. J. Intell. Syst..

[12]  John E. Laird,et al.  Learning General Search Control from Outside Guidance , 1987, IJCAI.

[13]  David Chapman,et al.  Planning for Conjunctive Goals , 1987, Artif. Intell..

[14]  Jaime G. Carbonell,et al.  Learning effective search control knowledge: an explanation-based approach , 1988 .

[15]  Prasad Tadepalli,et al.  Lazy ExplanationBased Learning: A Solution to the Intractable Theory Problem , 1989, IJCAI.

[16]  Manuela Veloso Nonlinear problem solving using intelligent casual-commitment , 1989 .

[17]  Jaime G. Carbonell,et al.  Learning by experimentation: the operator refinement method , 1990 .

[18]  David A. McAllester,et al.  Systematic Nonlinear Planning , 1991, AAAI.

[19]  Oren Etzioni,et al.  Integrating Abstraction and Explanation-Based Learning in PRODIGY , 1991, AAAI.

[20]  Ingrid Zukerman,et al.  Learning Search Control Rules for Planning: An Inductive Approach , 1991, ML.

[21]  Caroline Clarke Hayes Machining planning: a model of an expert level planning process , 1991 .

[22]  Subbarao Kambhampati,et al.  Explanation-Based Generalization of Partially Ordered Plans , 1991, AAAI.

[23]  Oren Etzioni,et al.  DYNAMIC: A New Role for Training Problems in EBL , 1992, ML.

[24]  Oren Etzioni,et al.  Why EBL Produces Overly-Specific Knowledge: A Critique of the PRODIGY Approaches , 1992, ML.

[25]  Manuela Veloso,et al.  An analysis of search techniques for a totally-ordered nonlinear planner , 1992 .

[26]  David Ruby,et al.  Learning Episodes for Optimization , 1992, ML.

[27]  Manuela Veloso Learning by analogical reasoning in general problem-solving , 1992 .

[28]  Oren Etzioni,et al.  PRODIGY4.0: The Manual and Tutorial , 1992 .

[29]  Yolanda Gil,et al.  Acquiring domain knowledge for planning by experimentation , 1992 .

[30]  Qiang Yang,et al.  Theory and Algorithms for Plan Merging , 1992, Artif. Intell..

[31]  Neeraj Bhatanagar Learning by incomplete explanations of failures in recursive domains , 1992, ICML 1992.

[32]  Qiang Yang,et al.  Handling feature interactions in process-planning , 1992, Appl. Artif. Intell..

[33]  Tom M. Mitchell,et al.  Learning by experimentation: acquiring and refining problem-solving heuristics , 1993 .

[34]  Raymond J. Mooney,et al.  Combining FOIL and EBG to Speed-up Logic Programs , 1993, IJCAI.

[35]  Manuela Veloso,et al.  Goal Interactions and Plan Quality , 1993 .

[36]  Mark A. Peot,et al.  Threat-Removal Strategies for Partial-Order Planning , 1993, AAAI.

[37]  M. A Perez,et al.  Automated Acquisition of Control Knowledge to Improve the Quality of Plans , 1993 .

[38]  Oren Etzioni,et al.  A Structural Theory of Explanation-Based Learning , 1993, Artif. Intell..

[39]  Xuemei Wang,et al.  Learning Planning Operators by Observation and Practice , 1994, AIPS.

[40]  Manuela M. Veloso,et al.  The Need for Different Domain-independent Heuristics , 1994, AIPS.

[41]  Masahiko Iwamoto A Planner with Quality Goal and Its Speed-up Learning for Optimization Problem , 1994, AIPS.

[42]  Yolanda Gil,et al.  Applying a General-Purpose Planning and Learning Architecture to Process Planning * , 1994 .

[43]  Eugene Fink,et al.  Prodigy Planning Algorithm , 1994 .

[44]  Manuela M. Veloso,et al.  Incremental Learning of Control Knowledge for Nonlinear Problem Solving , 1994, ECML.

[45]  Craig A. Knoblock Automatically Generating Abstractions for Planning , 1994, Artif. Intell..

[46]  Manuela M. Veloso,et al.  Linkability: Examining Causal Link Commitments in Partial-order Planning , 1994, AIPS.

[47]  Eugene Fink,et al.  Formalizing the PRODIGY planning algorithm , 1996 .