Chaotic Monkey Algorithm And Its Applications

Intelligent algorithm have been studied for a long time, and a lot of goodintelligence search algorithms have been introduced, including genetic algorithms(GA), particle swarm optimization (PSO) and artificial neural network (ANN)and so on. These classic intelligent algorithm substantially increase the humanability to solve complex problems, and they have been widely applied in a varietyof engineering projects. However, because almost all of these classical algorithmshard to escape the"curse of dimension", there still exit some restrictions whenputting them into implement.This thesis first introduces a new global optimization intelligent algorithm—monkey algorithm (MA). MA simulates the whole mountain climbing process ofmonkeys in nature, and designs three processes, namely climbing, watching, andjumping, to search for the global optimal solutions to continuous optimizationproblems. The algorithm is characterized by its insensitiveness to dimension ofthe optimization problem. The testing results of 11 functions show MA has theability to solve large-scale, multi-peak optimization problems, and enjoys quickspeed and high accuracy.Chaotic search, based on chaotic evaluation of variables, enjoys certainty, er-godicity and stochastic property, and has shown enormous ability of local search.In this thesis, a Chaotic Monkey Algorithm (CMA), which combines MA withchaotic search, is proposed. A set of sixteen well-known test functions is solvedwith CMA, and comparisons with MA, GA and PSO are made which show thatCMA can e?ectively enhance the searching effciency and greatly improve thesearching quality. Last but never least, the dimension test shows that CMA isinsensitive with the dimension of problems.In the end, CMA was used to solve a new kind of fuzzy constraint satisfac-tion problem. This thesis studies a kind of fuzzy constraint satisfaction problem(FCSP), in which the parameters with uncertainty are represented as fuzzy vari-ables. Using the credibility measure metrics the possibility of constraints, taking the joint credibility of all constraints as objective function, we transform theFCSP into an unconstrained optimization problem. A fuzzy simulation is usedto estimate the credibilities of the fuzzy events in the FCSP, based on which annew monkey algorithm is designed to solve the FCSP. Finally, several examplesare provided to illustrate the feasibility and the e?ectiveness of the proposedalgorithm.