Many-Objective Optimisation

Many-Objective Optimisation (MaO) refers to an optimisation problem having four or more objectives/criteria. With modern requirements, considering > 3 objectives in optimisation process has become increasingly common. Unfortunately, the increase in objective dimensionality brings big challenge, e.g., ineffectiveness of the Pareto dominance relation. My research interests in this regard include making conventional Pareto-based algorithms workable on many-objective problems and designing effective algorithms specifically for many-objective optimisation.

• Selected work

• [TEVC14] A little modification of density estimation in classic Pareto-based algorithms to make them work very well for many-objective optimisation problems. [Read More]

• [TEVC13, GECCO10] An evolutionary algorithm to explore the potential of the grid in solving many-objective problems. [Read More]

• [AIJ15] Converting a given many-objective optimisation problem into a bi-goal optimisation problem regarding convergence and diversity, and then handling it using the Pareto dominance relation in this bi-goal domain.

• [TEVC17] A vector angle-based many-objective evolutionary algorithm, which shares the similar idea as the decomposition-based method, but without the need of a set of weight vectors.