Explorations on template-directed genetic repair using ancient ancestors and other templates

Handling constraints for combinatorial optimization problems is a classic challenge faced by genetic and evolutionary algorithms. This paper explores a naturally inspired genetic repair process to enforce constraints on evolutionary search. Lolle et al. (2005) controversially claim that the model plant Arabidopsis thaliana appears to repair genetic errors using information inherited from ancestors other than the immediate parents [10] (i.e. non-Mendelian inheritance). We adapt this natural template-driven genetic repair process (GeneRepair) to help solve constraint problems. Building upon previous results [6][7][8] this paper explores repair templates that originate across a range of ancestors, between one and many thousands of generations old. The fitness of resulting populations are presented and compared to a benchmark technique using a random repair template [9]. The results show that very ancient (ancestral) repair templates perform best for larger problems, significantly outperforming the benchmark. The impact of background mutation rates on solution quality is also explored. Results suggest that ancestral repair is a good general-purpose constraint handling technique - helping to explain why this strategy might have evolved in nature.

[1]  Tim Walters,et al.  Repair and Brood Selection in the Traveling Salesman Problem , 1998, PPSN.

[2]  Amy FitzGerald,et al.  Genetic Repair for Optimization under Constraints Inspired by Arabidopsis Thaliana , 2008, PPSN.

[3]  Amy FitzGerald,et al.  Biologically inspired non-mendelian repair for constraint handling in evolutionary algorithms , 2010, GECCO '10.

[4]  J. Langdale,et al.  Faculty Opinions recommendation of Genome-wide non-mendelian inheritance of extra-genomic information in Arabidopsis. , 2005 .

[5]  Daniel Lichtblau.  Differential Evolution in Discrete and Combinatorial Optimization Mathematica tutorial notes , 2010 .

[6]  Lawrence Davis,et al.  Shall We Repair? Genetic AlgorithmsCombinatorial Optimizationand Feasibility Constraints , 1993, ICGA.

[7]  M. F. Fuller,et al.  Practical Nonparametric Statistics; Nonparametric Statistical Inference , 1973 .

[8]  D. O'Donoghue,et al.  GeneRepair - A Repair Operator for Genetic Algorithms , 2003 .

[9]  Xinyu Liu,et al.  Genetic Repair Strategies Inspired by Arabidopsisthaliana , 2009, AICS.

[10]  Chang Wook Ahn,et al.  A genetic algorithm for shortest path routing problem and the sizing of populations , 2002, IEEE Trans. Evol. Comput..

[11]  Takeshi Yamada,et al.  Conventional Genetic Algorithm for Job Shop Problems , 1991, ICGA.

[12]  W. J. Conover,et al.  Practical Nonparametric Statistics , 1972 .

[13]  Dr. Zbigniew Michalewicz,et al.  How to Solve It: Modern Heuristics , 2004 .

[14]  R. A. Groeneveld,et al.  Practical Nonparametric Statistics (2nd ed). , 1981 .

[15]  Gerhard Reinelt,et al.  TSPLIB - A Traveling Salesman Problem Library , 1991, INFORMS J. Comput..

[16]  Alice E. Smith,et al.  A genetic approach to the quadratic assignment problem , 1995, Comput. Oper. Res..

[17]  Carlos A. Coello Coello,et al.  THEORETICAL AND NUMERICAL CONSTRAINT-HANDLING TECHNIQUES USED WITH EVOLUTIONARY ALGORITHMS: A SURVEY OF THE STATE OF THE ART , 2002 .

[18]  Sancho Salcedo-Sanz,et al.  A survey of repair methods used as constraint handling techniques in evolutionary algorithms , 2009, Comput. Sci. Rev..

[19]  Thomas Bäck,et al.  A Comparative Study of a Penalty Function, a Repair Heuristic and Stochastic Operators with the Set-Covering Problem , 1995, Artificial Evolution.

[20]  A. Conejo,et al.  A parallel repair genetic algorithm to solve the unit commitment problem , 2003, 2003 IEEE Power Engineering Society General Meeting (IEEE Cat. No.03CH37491).