Experiments With Explicit ForLoops in Genetic RMIT University实验用显式循环遗传坎培门高等技术学院.ppt

Experiments With Explicit For-Loops in Genetic Programming Vic Ciesielski, Xiang Li {vc, xiali}@cs.rmit.edu.au School of Computer Science and Information Technology RMIT University, Australia Overview Introduction Why do we experiment loops? Related works Who have used loops in GP before? Syntax and Semantic of the For-Loops Experiments and Results Santa Fe Ant Problem Sorting Problem Conclusion Use loops to achieve your goals quicker and better Introduction Why do we experiment loops? Powerful – provide a mechanism for repetition. Why loops are seldom used in GP? Hard to evolve {start, end, body, update branch} Avoid infinity Large class of problems can be solved without loops Implicit loops are used. {Robsoccer, Santa Fe Ant} Type of loops For-loops and while-loops The research is focusing on for-loops. Loop Syntax Expectations Two for-loop constructs are used (FOR-LOOP1 NUM-INTERATIONS BODY) (FOR-LOOP2 START END BODY) Two strategies for setting values for NUM-ITERATIONS, START and END Set the value to a random integer (simple loops). Set the value to a result of any computation (unrestricted loops). Investigation and Expectation Convergence behavior, size and solutions Simple loops are easier to evolve than unrestricted loops Related Works [1] John R. Koza, Forrest H Bennet III, David Andre, and Martin A. Keane. Genetic Programming III; Darwinian invention and problem solving. Morgan Kaufmann, 1999. ADL – automatically defined loops – pre-established maximum number of execution [2] Kenneth E. Kinnear, Jr. Generality and difficulty in genetic programming: Evolving a sort. In Stephanie Forrest, editor, Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93,pages 287–294, University of Illinois at Urbana-Champaign, 17-21 1993.Morgan Kaufmann. A sorting algorithm – restricted total and individual loops times. – Inverse sized fitness measure [3] William B. Langdon. Data structures and genetic programming. In Peter J. Angeline and K. E. K