A Decomposition Based Evolutionary Algorithm for Many Objective Optimization with Systematic Sampling and Adaptive Epsilon Control

• Many objective optimization typically refers to problems with the number of objectives greater than four. • The commonly used dominance based methods for multi-objective optimization, such as NSGA-II, SPEA2 etc. are known to be inefficient for many-objective optimization as non-dominance does not provide adequate selection pressure to drive the population towards convergence. • There are also radically different approaches to deal with many objective optimization, such as attempts to identify the reduced set of objectives or corners of the Pareto front. Interactive use of decision makers preferences. • Use of reference points from systematic sampling or solution of the problem as a hypervolume maximization problem.