TY - CHAP

T1 - A novel hybrid algorithm for function optimization

T2 - Particle swarm assisted incremental evolution strategy

AU - Mo, W.

AU - Guan, S. G.

AU - Puthusserypady, Sadasivan K.

PY - 2007

Y1 - 2007

N2 - This chapter presents a new algorithm for function optimization problems, particle swarm assisted incremental evolution strategy (PIES), which is designed for enhancing the performance of evolutionary computation techniques by evolving the input variables incrementally. The whole evolution consists of several phases and one more variable is focused in each phase. The number of phases is equal to the number of variables in maximum. Each phase is composed of two stages: In the single-variable evolution (SVE) stage, a population is evolved with respect to one independent variable in a series of cutting planes; in the multivariable evolving (MVE) stage, the initial population is formed by integration. The integration integrates the solutions found by the SVE stage in the current phase and the solutions found by the MVE stage in the last phase. Subsequently the population is evolved with respect to the incremented variable set in a series of cutting hyperplanes. To implement this incremental optimization, a combination of evolution strategy (ES) and particle swarm optimization (PSO) is used. ES is applied to searching optima in the cutting planes/hyperplanes, while PSO is applied to adjust the cutting planes (in SVE stages) or hyperplanes (in MVE stages). The experiment results show that PIES generally outperforms three other evolutionary algorithms, improved normal GA, PSO, and SADE_CERAF, in the sense that PIES can find solutions closer to the true optima both in the variable space and in the objective space.

AB - This chapter presents a new algorithm for function optimization problems, particle swarm assisted incremental evolution strategy (PIES), which is designed for enhancing the performance of evolutionary computation techniques by evolving the input variables incrementally. The whole evolution consists of several phases and one more variable is focused in each phase. The number of phases is equal to the number of variables in maximum. Each phase is composed of two stages: In the single-variable evolution (SVE) stage, a population is evolved with respect to one independent variable in a series of cutting planes; in the multivariable evolving (MVE) stage, the initial population is formed by integration. The integration integrates the solutions found by the SVE stage in the current phase and the solutions found by the MVE stage in the last phase. Subsequently the population is evolved with respect to the incremented variable set in a series of cutting hyperplanes. To implement this incremental optimization, a combination of evolution strategy (ES) and particle swarm optimization (PSO) is used. ES is applied to searching optima in the cutting planes/hyperplanes, while PSO is applied to adjust the cutting planes (in SVE stages) or hyperplanes (in MVE stages). The experiment results show that PIES generally outperforms three other evolutionary algorithms, improved normal GA, PSO, and SADE_CERAF, in the sense that PIES can find solutions closer to the true optima both in the variable space and in the objective space.

UR - http://www.scopus.com/inward/record.url?scp=34548303587&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-73297-6_5

DO - 10.1007/978-3-540-73297-6_5

M3 - Chapter

AN - SCOPUS:34548303587

SN - 3540732969

SN - 9783540732969

T3 - Studies in Computational Intelligence

SP - 101

EP - 125

BT - Hybrid Evolutionary Algorithms

A2 - Grosan, Crina

A2 - Abraham, Ajith

A2 - Ishibuchi, Hisao

ER -