An Adaptive Opposition-Based Learning Selection: The Case for Jaya Algorithm

Over the years, opposition-based Learning (OBL) technique has been proven to effectively enhance convergence of meta-heuristic algorithms. The fact that OBL is able give alternative candidate solutions in one or more opposite directions ensures good exploration and exploitation search space. In last decade, many techniques have established literature including Standard-OBL, General-OBL, Quasi Reflection-OBL, Centre-OBL Optimal-OBL. Although useful, much existing adoption into algorithms based on a single technique. If space contains peaks with potentially local optima, relying may not be sufficiently effective. fact, if are close together, prevent entrapment optima. Addressing this issue, assembling sequence algorithm can useful overall performance. Based simple penalized reward mechanism, best performing rewarded continue its execution next cycle, whilst poor will miss cease current turn. This paper presents new adaptive approach integrating than Jaya Algorithm, termed OBL-JA. Unlike other adoptions which use type OBL, OBL-JA uses several OBLs their selections each individual Experimental results using combinatorial testing problems as case study demonstrate shows very competitive against works term test suite size. also show performs better standard Algorithm most tested cases due ability adapt behaviour performance feedback process.