NATURE INSPIRED COMPUTATION METHODS AND THEIR APPLICATION IN FUNCTION OPTIMIZATION

Abstract

The thesis entitled Nature Inspired Computation (NIC):Methods and their application in function optimization is intended to present the state of art Nature Inspired Computing (NIC) techniques and their applications in the field of optimization. Here this concern research focuses on three well known Nature Inspired Algorithms (NIAs). newlineFirstly, it focuses on the Artificial Bee Colony algorithm. Here this thesis suggest some modified hybrids of basic ABC algorithm like, Randomized Memetic ABC (RMABC) by adding two new parameters in Memetic ABC, improved onlooker bee phase in ABC (IoABC), enhanced local search in ABC (EnABC), improved Memetic search in ABC (IMeABC), fitness based position update in ABC (FPABC), Memetic search in FPABC (MFPABC), new local search strategy in ABC (NLSSABC) by introducing a new local search phase on ABC inspired by golden section search and a hybrid of levy flight search and Memetic search strategy in ABC (LFMABC). Secondly this thesis concerns with Differential Evolution (DE) an evolutionary algorithm. This thesis suggests three modifications in basic DE. First, Memetic search inspired by golden section search incorporated in basic DE (MSDE). Second, levy flight search in fitness based DE (LFBDE). Third, opposition based levy flight search in DE (OLFDE). Finally, it focuses on newly developed population based nature inspired algorithm named Spider Monkey Optimization (SMO). This thesis suggests three modifications in basic SMO algorithm.This thesis suggests three modifications in basic SMO algorithm. First, it proposed a modified position update in SMO (MPU-SMO) algorithm by enhancing position update strategy in both local leader phase and global leader phase. Second, a fitness based position update in SMO (FPSMO) and third opposition based learning in SMO (OBSMO) algorithm. In order to establish superiority of proposed algorithms over basic algorithms and their recent variants, they are tested over a set of benchmark functions and some real world optimization problems.