Hybridisations of Variable Neighbourhood and Modified Simplex Methods to Harmony Search Algorithm for Process Optimisations

摘要

Nowadays, the engineering problem systems are large and complicated. An effective finite sequence of instructions for solving these problems can be categorised into optimisation and meta-heuristic algorithms. Though the best decision variable levels from some sets of available alternatives cannot be done, meta-heuristics is an alternative for experience-based techniques that rapidly help in problem solving, learning and discovery in the hope of obtaining a more efficient or more robust procedure. All meta-heuristics provide auxiliary procedures in terms of their own tooled box functions. It has been shown that the effectiveness of all meta-heuristics depends almost exclusively on these auxiliary functions. In fact, the auxiliary procedure from one can be implemented into other meta-heuristics. A well-known meta-heuristic of harmony search algorithm (HSA) is compared with its hybridisations. HSA is used to produce a near optimal solution under a consideration of the perfect state of harmony of the improvisation process of musicians. This study presents solution procedures via constrained and unconstrained problems with different natures of single and multi peak surfaces including a curved ridge surface. HSA is modified via variable neighbourhood search method (VNSM) philosophy including a modified simplex method (MSM). The basic idea is the change of neighbourhoods during searching for a better solution. The hybridisations proceed by a descent method to a local minimum exploring then, systematically or at random, increasingly distant neighbourhoods of this local solution. The results show that a variant of HSA with VNSM and MSM seems to be better in terms of the mean and variance of design points and yields.