Surrogate-assisted multicriteria optimization: Complexities, prospective solutions, and business case

Abstract

Complexity in solving real-world multicriteria optimization problems often stems from the fact that complex, expensive, and/or time-consuming simulation tools or physical experiments are used to evaluate solutions to a problem. In such settings, it is common to use efficient computational models, often known as surrogates or metamodels, to approximate the outcome (objective or constraint function value) of a simulation or physical experiment. The presence of multiple objective functions poses an additional layer of complexity for surrogate-assisted optimization.For example, complexities may relate to the appropriate selection of metamodels for the individual objective functions, extensive training time of surrogate models, or the optimal use of many-core computers to approximate efficiently multiple objectives simultaneously. Thinking out of the box, complexity can also be shifted from approximating the individual objective functions to approximating the entire Pareto front. This leads to further complexities, namely, how to validate statistically and apply the techniques developed to real-world problems. In this paper, we discuss emerging complexity-related topics in surrogate-assisted multicriteria optimization that may not be prevalent in nonsurrogate-assisted single-objective optimization. These complexities are motivated using several real-world problems in which the authors were involved. We then discuss several promising future research directions and prospective solutions to tackle emerging complexities in surrogate-assisted multicriteria optimization. Finally, we provide insights from an industrial point of view into how surrogate-assisted multicriteria optimization techniques can be developed and applied within a collaborative business environment to tackle real-world problems.

Weighted ensembles in model-based global optimization

“…The convex model combinations presented in this paper can be viewed as an elegant stacking approach and as such they are similar to 'ensembles of surrogates', which however use a fixed rule for determining weights. Dynamical model selection can also be achieved by reinforcement learning [13], which would be an alternative for single model selection by optimization ('Choose'). Furthermore, it will be interesting to extend the set of test problems and include more types of base models.…”

Surrogate-Assisted Evolutionary Optimization of Large Problems

“…The sample selection for MOPs is more complicated than for SOPs, because one needs to consider both convergence and diversity [2]. In the literature, several approaches have been used, such as selecting a set of uniformly distributed samples in the objective space [3,9,33] or a set of isolated samples in the decision space [1,36], and using ExI [55], LCB [38] or expected hypervolume improvement [37,42].…”

A Random Forest-Assisted Evolutionary Algorithm for Data-Driven Constrained Multiobjective Combinatorial Optimization of Trauma Systems

“…Most existing SAEAs have been developed for online datadriven optimization of continuous problems, where appropriate surrogates are chosen and model management strategies are designed to make sure that the EA is able to find the best solution with the given computation budget [12]. A wide range of regression models have been adopted as the surrogates, such as Kriging model (Gaussian processes regression model) [13], [14], radial basis function (RBF) networks [15]- [18], polynomial regression [19], and artificial neural networks [20], [21].…”

“…In individualbased model management strategies, promising and uncertain solutions according to the surrogate model [26]- [28] are evaluated using the expensive objective functions. A large number of SAEAs for single-objective optimization [6], [25], [29], [30], multi-objective optimization [12], [22], [31], [32], and many-objective optimization [13], [33] have been proposed.…”