A swarm with an effective information sharing mechanism for unconstrained and constrained single objective optimisation problems

Abstract

We present an effective multilevel information sharing strategy within a swarm to handle single objective, constrained and unconstrained optimization problems. A swarm is considered as a collection of individuals having a common goal to reach the best value (minimum or maximum) of a function. The success of a swarm is attributed to the identification of a set of competent leaders and a meaningful information sharing scheme between the leaders and the rest of the individuals that enables the swarm to collectively attain the common goal. The proposed algorithm mimics the above behavioral processes of a real swarm and maintains unique individuals at all time instants. The uniqueness among the individuals result in a set of near optimal solutions at the final phase that is useful for sensitivity analysis. The benefits of the effective information sharing strategy is illustrated by solving two unconstrained problems with multiple equal and unequal optima and a constrained optimization problem.