Application of cohort intelligence algorithm for goal programming problems with improved constraint handling method

Goal programming (GP) is a satisficing-based mathematical modelling technique. In this paper, cohort intelligence (CI) algorithm and its variations are applied to solve a variety of GP problems. The penalty function-based and probability-based constrained handling approaches are applied. Furthermore, a hybridisation of PF and prob-based approaches is developed to handle hard constraints effectively. The proposed approach is validated by solving five benchmark problems as well as practically important real-world truss design, welding beam design, metal cutting, supplier selection, capital budgeting, and staff scheduling problems. The solutions are compared with evolutionary algorithms and LINGO. The results obtained are exceedingly better in terms of satisfying the hard constraints as well as minimising the deviations from the set goals. It is important to note that for truss design, metal cutting and supplier selection problems, all the hard constraints are satisfied using the proposed technique as against with the SA, PSO & Tabu Search.