[00070] SDDS-SABC based Algorithm for solving non-linear optimization problems
Session Time & Room : 3C (Aug.23, 13:20-15:00) @D502
Type : Contributed Talk
Abstract : Optimizing complex non-linear constrained optimization problems is often a challenging task. This work proposes a new hybrid method called SDDS-SABC based on the Split-Detect-Discard-Shrink technique and Sophisticated ABC algorithm to optimize the said problems. The SDDS method is responsible for shrinking the full search region through a recursive breakdown and improves computational effort to focus on the subregion covering potential solutions for further decomposition. SABC plays a vital role in extracting the best solutions from the subregions whose values help detect the promising subregion. Both SDDS and SABC are sequentially repeated until the region reduces to a nominal width representing the optimization problem's global/ close to global solution(s). The Ranking and selection rules have been applied to assist optimistic decision-making with an attitude to discard the subregion covering non-promising solution (s). At the same time, the subregion with a promising solution is accepted as the current shrink region for a further split. We introduce a new initialization scheme for food sources in the SABC algorithm, which excels the existing initialization process. Develop Dual-strategy Employed bee's phase, allowing bees to split into two groups and use their respective group strategies to explore their neighbourhood while maintaining their collaborative contribution. We also introduce a new Dynamic penalty method that is free from extra parameters or factors like most existing penalty methods do to improve the optimization efficiency. To check the validity of SDDS-SABC, we have applied it to benchmark functions and engineering problems. To measure our proposed method's statistical significance against other existing heuristic optimization methods, we carried out the non-parametric Friedman and Wilcoxon rank tests.
