Leonardo C. T. Bezerra

Automatic Component-Wise Design of Multiobjective Evolutionary Algorithms

Multiobjective evolutionary algorithms (MOEAs) are typically proposed, studied, and applied as monolithic blocks with a few numerical parameters that need to be set. Few works have studied how the algorithmic components of these evolutionary algorithms can be classified and combined to produce new algorithmic designs. The motivation for studies of this latter type stem from the development of flexible software frameworks and the usage of automatic algorithm configuration methods to find novel algorithm designs. In this paper, we propose an MOEA template and a new conceptual view of its components that surpasses existing frameworks in both number of algorithms that can be instantiated from the template and flexibility to produce novel algorithmic designs. We empirically demonstrate the flexibility of our proposed framework by automatically designing MOEAs for continuous and combinatorial optimization problems. The automatically designed algorithms are often able to outperform six traditional MOEAs from the literature, even after tuning their numerical parameters.

Automatic Design of Evolutionary Algorithms for Multi-Objective Combinatorial Optimization

Multi-objective evolutionary algorithms (MOEAs) have been the subject of a large research effort over the past two decades. Traditionally, these MOEAs have been seen as monolithic units, and their study was focused on comparing them as blackboxes. More recently, a component-wise view of MOEAs has emerged, with flexible frameworks combining algorithmic components from different MOEAs. The number of available algorithmic components is large, though, and an algorithm designer working on a specific application cannot analyze all possible combinations. In this paper, we investigate the automatic design of MOEAs, extending previous work on other multi-objective metaheuristics. We conduct our tests on four variants of the permutation flowshop problem that differ on the number and nature of the objectives they consider. Moreover, given the different characteristics of the variants, we also investigate the performance of an automatic MOEA designed for the multi-objective PFSP in general. Our results show that the automatically designed MOEAs are able to outperform six traditional MOEAs, confirming the importance and efficiency of this design methodology.

Deconstructing Multi-objective Evolutionary Algorithms: An Iterative Analysis on the Permutation Flow-Shop Problem

Many studies in the literature have applied multi-objective evolutionary algorithms (MOEAs) to multi-objective combinatorial optimization problems. Few of them analyze the actual contribution of the basic algorithmic components of MOEAs. These components include the underlying EA structure, the fitness and diversity operators, and their policy for maintaining the population. In this paper, we compare seven MOEAs from the literature on three bi-objective and one tri-objective variants of the permutation flowshop problem. The overall best and worst performing MOEAs are then used for an iterative analysis, where each of the main components of these algorithms is analyzed to determine their contribution to the algorithms’ performance. Results confirm some previous knowledge on MOEAs, but also provide new insights. Concretely, some components only work well when simultaneously used. Furthermore, a new best-performing algorithm was discovered for one of the problem variants by replacing the diversity component of the best performing algorithm (NSGA-II) with the diversity component from PAES.

Comparing Decomposition-Based and Automatically Component-Wise Designed Multi-Objective Evolutionary Algorithms

A main focus of current research on evolutionary multi-objective optimization (EMO) is the study of the effectiveness of EMO algorithms for problems with many objectives. Among the several techniques that have led to the development of more effective algorithms, decomposition and component-wise design have presented particularly good results. But how do they compare? In this work, we conduct a systematic analysis that compares algorithms produced using the MOEA/D decomposition-based framework and the AutoMOEA component-wise design framework. In particular, we identify a version of MOEA/D that outperforms the best known MOEA/D algorithm for several scenarios and confirms the effectiveness of decomposition on problems with three objectives. However, when we consider problems with five objectives, we show that MOEA/D is unable to outperform SMS-EMOA, being often outperformed by it. Conversely, automatically designed AutoMOEAs display competitive performance on three-objective problems, and the best and most robust performance among all algorithms considered for problems with five objectives.

Automatic generation of multi-objective ACO algorithms for the bi-objective knapsack

Multi-objective ant colony optimization (MOACO) algorithms have shown promising results for various multi-objective problems, but they also offer a large number of possible design choices. Often, exploring all possible configurations is practically infeasible. Recently, the automatic configuration of a MOACO framework was explored and was shown to result in new state-of-the-art MOACO algorithms for the bi-objective traveling salesman problem. In this paper, we apply this approach to the bi-objective bidimensional knapsack problem (bBKP) to prove its generality and power. As a first step, we tune and improve the performance of four MOACO algorithms that have been earlier proposed for the bBKP. In a second step, we configure the full MOACO framework and show that the automatically configured MOACO framework outperforms all previous MOACO algorithms for the bBKP as well as their improved variants.

An Empirical Assessment of the Properties of Inverted Generational Distance on Multi- and Many-Objective Optimization

The inverted generational distancei¾źIGD is a metric for assessing the quality of approximations to the Pareto front obtained by multi-objective optimization algorithms. The IGD has become the most commonly used metric in the context of many-objective problems, i.e., those with more than three objectives. The averaged Hausdorff distance and

An analysis of local search for the bi-objective bidimensional knapsack problem

\textit{IGD}^+

Automatic Generation of MOACO Algorithms for the Biobjective Bidimensional Knapsack Problem

are variants of the IGD proposed in order to overcome its major drawbacks. In particular, the IGD is not Pareto compliant and its conclusions may strongly change depending on the size of the reference front. It is also well-known that different metrics assign more importance to various desired features of approximation fronts, and thus, they may disagree when ranking them. However, the precise behavior of the IGD variants is not well-understood yet. In particular,