Performance of a steady state quantum genetic algorithm for multi/many-objective engineering optimization problems

In this paper, we introduce a novel decomposition based steady state quantum genetic algorithm for the solution of engineering optimization problems. Systematic sampling is used to generate reference directions and a small population of quantum individuals (solutions with variables represented as Q-bits) is evolved using a simple variation operator. A solution represented using Q-bits has the ability to probabilistically represent a number of solutions defined through observation. We exploit the benefits of quantum representation within a steady state evolution scheme and illustrate the behavior of the algorithm using unconstrained DTLZ2 test problem involving 2, 3, 5, and 8 objectives and a set of multi/many-objective constrained engineering design optimization problems. The underlying motivation of quantum representation stems from its ability to represent multiple states which offers the potential to evolve a small population of solutions. This aspect is magnified even further when one attempts to solve a many objective optimization problem where evolution of a large population of solutions may not be practically viable. The proposed approach is expected to gain more attention in near future as quantum computing infrastructures become more readily available.