Using genetic programming to learn and improve control knowledge.

DOI10.1016/S0004-3702(02)00246-1MaRDI QIDQ1852856zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Ricardo Aler, Daniel Borrajo, Pedro Isasi

Publication date 21 January 2003

Published in Artificial Intelligence (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/s0004-3702(02)00246-1, http://hdl.handle.net/10016/3937

zbMATH Keywords

Genetic programmingMulti-strategy learningPlanningSpeedup learning

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Review text

The purpose of this article is to present a multi-strategy approach to learn heuristics for planning. This multi-strategy system, called Hamlet-EvoCK, combines a learning algorithm specialized in planning (Hamlet) and a Genetic Programming (GP) based system (EvoCK: Evolution of Control Knowledge). Both systems are able to learn heuristics for planning on their own, but both of them have weaknesses. Based on previous experience and some experiments performed in this article, it is hypothesized that Hamlet handicaps are due to its example-driven operators and not having a way to evaluate the usefulness of its control knowledge. It is also hypothesized that even if Hamlet control knowledge is sometimes incorrect, it might be easily correctable. For this purpose, a GP-based stage is added, because of its complementary biases: GP genetic operators are not example-driven and it can use a fitness function to evaluate control knowledge. Hamlet and EvoCK are combined by seeding EvoCK initial population with Hamlet control knowledge. It is also useful for EvoCK to start from a knowledge-rich population instead of a random one. By adding the GP stage to Hamlet, the number of solved problems increases from 58\% to 85\% in the blocks world and from 50\% to 87\% in the logistics domain (0\% to 38\% and 0\% to 42\% for the hardest instances of problems considered).

Recommendations

• Evolving programs and solutions using genetic programming with application to learning and adaptive control
• scientific article; zbMATH DE number 4106289
• Genetic learning of fuzzy controllers
• Learning fuzzy rules for controllers with genetic algorithms
• AN IMPROVED KNOWLEDGE-ACQUISITION STRATEGY BASED ON GENETIC PROGRAMMING
• Genetic search-based control

Cites work

• scientific article; zbMATH DE number 3497315 (Why is no real title available?)
• scientific article; zbMATH DE number 795579 (Why is no real title available?)
• scientific article; zbMATH DE number 795587 (Why is no real title available?)
• scientific article; zbMATH DE number 1448981 (Why is no real title available?)
• Integrating planning and learning: the PRODIGY architecture
• Macro-operators: A weak method for learning

Cited in

(4)

• Heuristic learning based on genetic programming
• Investigation of different seeding strategies in a genetic planner
• AN IMPROVED KNOWLEDGE-ACQUISITION STRATEGY BASED ON GENETIC PROGRAMMING
• scientific article; zbMATH DE number 1962033 (Why is no real title available?)

Describes a project that uses

Uses Software

• Graphplan
• Prodigy
• UCPOP
• Walksat

This page was built for publication: Using genetic programming to learn and improve control knowledge.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1852856)