Synergy of ICA and MCDM for multi‐response optimisation problems

Of late, attempts are being made to optimise production system problems by minimum cost. A good available device in this area is response surface methodology. This methodology combines experimental designs and statistical techniques for empirical model building and optimising. In most situations simulated models for real world problems are non‐linear multi‐response, while responses are conflicting. The simultaneous optimisation of several conflicting responses is computationally expensive. So this makes the problem solving extremely complex. Since few attempts have been made to scrutinise this domain, in this paper the nonlinear continuous multi‐response problem is investigated. In order to tackle multi‐response optimisation difficulties, we propose a new hybrid meta heuristic based on the imperialist competitive algorithm. It simulates a socio–economical procedure, imperialistic competition. When there are some non‐dominated solutions in searching space, a technique for order performance by similarity to ideal solution is used to identify which non‐dominated solutions are imperialists and which ones belong to colonial societies. A particle swarm‐like mechanism is employed to model the influence of imperialists on colonies. The algorithm will continue until only one imperialist obtains all countries’ possessions. In order to prevent carrying out extensive experiments to find optimum parameters of the algorithm, we apply the Taguchi approach. Since there is no standard benchmark in this field, we use three case studies from distinguished papers in the multi‐response optimisation field. Comparing the results with some works mentioned in the literature reveals the superiority of the proposed algorithm.