FINCH is a methodology for evolving Java bytecode,enabling the evolution of extant, unrestricted Javaprograms, or programs in other languages that compileto Java bytecode. The established approach in geneticprogramming (GP) involves the definition of functionsand terminals appropriate to the problem at hand, afterwhich evolution of expressions using these definitionstakes place. FINCH evolutionarily improves actual,extant software, which was not intentionally writtenfor the purpose of serving as a GP representation inparticular, nor for evolution in general. In this workwe show how several game heuristics that are taken asreal-world Java programs are effortlessly andautomatically improved by FINCH. We have developed apowerful tool [1,2,3] by which extant software, writtenin the Java programming language, or in a language thatcompiles to Java bytecode, can be evolved directly,without an intermediate genomic representation, andwith no restrictions on the constructs used. We providean overview of this system, some previous results, itsusability, and the application of FINCH to evolvingJava-written game heuristics.
CITATION STYLE
Orlov, M., Bregman, C., & Sipper, M. (2011). Automatic Evolution of Java-Written Game Heuristics (pp. 277–277). https://doi.org/10.1007/978-3-642-23716-4_30
Mendeley helps you to discover research relevant for your work.