Evolutionary Large‐Scale Multi‐Objective Optimization

Large‐scale multi‐objective optimization problems (LSMOPs) refer to a subset of optimization problems which have both a large number of decision variables and multiple conflicting objectives. The sparse large‐scale multi‐objective optimization problems, namely, sparse LSMOPs, refer to a class of LSMOPs whose decision variables in Pareto optimal solutions are mostly zero‐valued. This chapter provides a brief introduction to evolutionary large‐scale multi‐objective Optimization. It introduces some primary concepts followed by the introduction of some commonly used large‐scale multi‐objective optimization test problems. The chapter, then, elaborates three kinds of performance indicators, namely, the indicators assessing only convergence, the indicators assessing only diversity, and the indicators assessing both convergence and diversity. It introduces the test problems that are tailored for sparse large‐scale multi‐objective optimization. Finally, a dedicated performance indicator which assesses convergence, sparsity, and diversity simultaneously is introduced in detail.