An Improved Cohort Intelligence with Panoptic Learning Behavior for Solving Constrained Problems

In this paper, we present a new optimization algorithm referred to as Cohort Intelligence with Panoptic learning (CI-PL). This proposed algorithm is a modified version of Cohort Intelligence (CI), where Panoptic learning (PL) is incorporated into CI which makes every cohort candidate learn the most from the best candidate but at same time it does not completely ignore the other candidates. The PL is assisted with a new sampling interval reduction method based on the standard deviation between the behaviors of the cohort candidates. A variety of well-known set of unconstrained and constrained test problems have been successfully solved by using the proposed algorithm. The CI-PL approach produced competent and sufficiently robust results solving unconstrained, constrained, and engineering problems. The associated strengths, weaknesses, and possible real-world extensions are also discussed.