PROJECT TITLE :

A Multiobjective Evolutionary Algorithm Using Gaussian Process-Based Inverse Modeling

ABSTRACT:

To approximate the Pareto front, most existing multiobjective evolutionary algorithms store the nondominated solutions found thus way within the population or in an external archive throughout the search. Such algorithms often need a high degree of diversity of the stored solutions and only a restricted range of solutions can be achieved. By contrast, model-based algorithms will alleviate the requirement on answer diversity and in principle, as many solutions as required will be generated. This paper proposes a new model-based mostly methodology for representing and searching nondominated solutions. The most idea is to construct Gaussian method-based mostly inverse models that map all found nondominated solutions from the target space to the choice space. These inverse models are then used to make offspring by sampling the objective house. To facilitate inverse modeling, the multivariate inverse operate is decomposed into a group of univariate functions, where the amount of inverse models is reduced using a random grouping technique. In depth empirical simulations demonstrate that the proposed algorithm exhibits sturdy search performance on a selection of medium to high dimensional multiobjective optimization check issues. Further nondominated solutions are generated a posteriori using the constructed models to increase the density of solutions in the popular regions at a low computational cost.