A Decomposition-Based Multiobjective Optimization Evolutionary Algorithm with Adaptive Weight Generation Strategy

Multiobjective evolutionary algorithm based on decomposition (MOEA/D) is the seminal framework of multiobjective evolutionary algorithms (MOEAs). To alleviate the nonuniformly distributed solutions generated by a fixed set of evenly distributed weight vectors in the presence of nonconvex and disconnected problems, an adaptive vector generation mechanism is proposed. A coevolution strategy and a vector generator are synergistically cooperated to remedy the weight vectors. Optimal weight vectors are generated to replace the useless weight vectors to assure that optimal solutions are distributed evenly. Experiment results indicate that this mechanism is efficient in improving the diversity of MOEA/D.

An enhanced auto adaptive vector evaluated-based metaheuristic for solving real-world problems

This article investigates the enhancement of a vector evaluat-ed-based adaptive metaheuristics for solving two multiobjective problems called environmental-economic dispatch and portfolio optimization. The idea is to evolve two populations independently, and exchange information between them, i.e., the first population evolves according to the best individual of the second population and vice-versa. The choice of which algorithm will be executed on each generation is carried out stochastically among three evolutionary algorithms well-known in the literature: PSO, DE, ABC. To assess the results, we used an established metric in multiobjective evolutionary algorithms called hypervolume. Tests solving the referred problem have shown that the new approach reaches the best hypervolumes in power systems comprised of six and forty generators and five different datasets of portfolio optimization. The experiments were performed 31 times, using 250, 500, and 1000 iterations in both problems. Results have also shown that our proposal tends to overcome a variation of a hybrid SPEA2 compared to their cooperative and competitive approaches.

A Hybrid Strategic Oscillation with Path Relinking Algorithm for the Multiobjective k-Balanced Center Location Problem

This paper presents a hybridization of Strategic Oscillation with Path Relinking to provide a set of high-quality nondominated solutions for the Multiobjective k-Balanced Center Location problem. The considered location problem seeks to locate k out of m facilities in order to serve n demand points, minimizing the maximum distance between any demand point and its closest facility while balancing the workload among the facilities. An extensive computational experimentation is carried out to compare the performance of our proposal, including the best method found in the state-of-the-art as well as traditional multiobjective evolutionary algorithms.

Resampled Efficient Frontier Integration for MOEAs

Mean-variance portfolio optimization is subject to estimation errors for asset returns and covariances. The search for robust solutions has been traditionally tackled using resampling strategies that offer alternatives to reference sets of returns or risk aversion parameters, which are subsequently combined. The issue with the standard method of averaging the composition of the portfolios for the same risk aversion is that, under real-world conditions, the approach might result in unfeasible solutions. In case the efficient frontiers for the different scenarios are identified using multiobjective evolutionary algorithms, it is often the case that the approach to averaging the portfolio composition cannot be used, due to differences in the number of portfolios or their spacing along the Pareto front. In this study, we introduce three alternatives to solving this problem, making resampling with standard multiobjective evolutionary algorithms under real-world constraints possible. The robustness of these approaches is experimentally tested on 15 years of market data.

Joint Structure and Parameter Optimization of Multiobjective Sparse Neural Network

This work addresses the problem of network pruning and proposes a novel joint training method based on a multiobjective optimization model. Most of the state-of-the-art pruning methods rely on user experience for selecting the sparsity ratio of the weight matrices or tensors, and thus suffer from severe performance reduction with inappropriate user-defined parameters. Moreover, networks might be inferior due to the inefficient connecting architecture search, especially when it is highly sparse. It is revealed in this work that the network model might maintain sparse characteristic in the early stage of the backpropagation (BP) training process, and evolutionary computation-based algorithms can accurately discover the connecting architecture with satisfying network performance. In particular, we establish a multiobjective sparse model for network pruning and propose an efficient approach that combines BP training and two modified multiobjective evolutionary algorithms (MOEAs). The BP algorithm converges quickly, and the two MOEAs can search for the optimal sparse structure and refine the weights, respectively. Experiments are also included to prove the benefits of the proposed algorithm. We show that the proposed method can obtain a desired Pareto front (PF), leading to a better pruning result comparing to the state-of-the-art methods, especially when the network structure is highly sparse.