Ant colony optimisation algorithm for multiobjective subset selection problems

Multiobjective subset selection problems widely exist in many real-world applications. Multiobjective ant colony optimisation (MOACO) is a strong and kind instrument for settling those issues. However, it still has two shortages which the authors must resolve. One, its solution construction process is inconsistent with the disorder characteristics of solutions, which prevent it from getting better solutions. Two, conventional MOACOs which most deal with biobjective optimisation problems are difficult to figure out high-dimensional objectives optimisation problems. In this study, the authors propose a new MOACO to resolve those two disadvantages. They give its detailed descriptions and an exhaust experiment. And the results on C measure, spacing measure and inverted generational distance (IGD) show that the proposed algorithm has a powerful convergence ability and gets a better balance between convergence and diversity compared with other state of art approaches.