A runnable functional formal memetic algorithm framework

Abstract

Historically Functional Programming FP for short has been associated with a small scope of applications mainly academic The computer science community did not pay enough attention to its potential perhaps due to the lack of e ciency of functional languages Now new theoretical developments in the eld of FP are emerging and better languages e g Haskell Concurrent and Parallel Haskell have been de ned and implemented Genetic algorithms GA are search and optimization techniques which work on a nature inspired principle the Darwinian evolution The corner idea of Darwin theory is that of natural selection The concept of natural selection is captured by GA Speci cally solutions to a given problem are codi ed in the so called chromosomes The evolution of chromosomes due to the action of crossover mutation and natural selection is simulated through computer code GA have been broadly applied and recognized as a robust search and optimization technique GA combined with a local search stage were called Memetic Algorithms after In this paper a functional framework for formal memetic algorithms is intro duced It can be easily extended by subclassi cation of the class hierarchy to provide genetic algorithm specialization memetic algorithm genetic algorithm with islands of possible solutions etc and additional genetic operators behavior To run the frame work over a particular problem a proper encoding of chromosomes should be provided with an instantiation of the genetic operators We claim that functional programming languages at least the one in which our framework has been developed Haskell have reached the necessary maturity to deal with combinatorial optimization problems