- Research Article
9
- 10.2139/ssrn.3792994
Explainable Interactive Evolutionary Multiobjective Optimization
- Jan 01, 2021
- SSRN Electronic Journal
- Salvatore Corrente + 3 more +3
Explainable Interactive Evolutionary Multiobjective Optimization
Recent studies show that evolutionary optimizers are effective tools in solving real-world problem with complex and competing specifications. As more advanced multiobjective evolutionary optimizers (MOEO) are being designed and proposed, the issue of performance assessment has become increasingly important. While performance assessment could be done via theoretical and empirical approach, the latter is more practical and effective and has been adopted as the de facto approach in the evolutionary multiobjective optimization community. However, researches pertinent to empirical study have focused mainly on its individual components like test metrics and functions, there are limited discussions on the overall adequacy of empirical test in substantiating their statements made on the performance and behavior of the evaluated algorithm. As such, this paper aims to provide a holistic perspective towards the empirical investigation of MOEO and present a conceptual framework, which researchers could consider in the design and implementation of MOEO experimental study. This framework comprises of a structural algorithmic development plan and a general theory of adequacy in the context of evolutionary multiobjective optimization.
Explainable Interactive Evolutionary Multiobjective Optimization
Explainable Interactive Evolutionary Multiobjective Optimization
Constructing evolutionary algorithms for bilevel multiobjective optimization
We propose a procedure to construct evolutionary bilevel optimization algorithms based on recent theoretical advances that have established connections between bilevel optimization and multiobjective optimization. In the proposed procedure, a new algorithm is defined by integrating an evolutionary multiobjective optimization algorithm with a partial order that is compatible with bilevel optimization. The advantages of the procedure include the ability to harness the methodology of evolutionary multiobjective optimization for bilevel optimization and to systematically develop new algorithms for single-objective and multiobjective bilevel optimization. No regularity assumptions are used, which ensures maximal applicability of the optimization algorithms constructed by the procedure. The necessary theoretical foundation is developed and the steps of the procedure are illustrated with an example.
Read moreEditorial for the Special Issue on Theoretical Aspects of Evolutionary Multi-Objective Optimization
Evolutionary algorithms have been widely used to tackle multi-objective optimization problems. Despite many successful applications there is a lack of theoretical knowledge on how and why this type of algorithm works particularly well for multi-objective problems. This special issue presents current advances to the theoretical understanding of evolutionary algorithms for multi-objective optimization. It comprises four very high quality papers which have been selected via a rigorous reviewing process. The first paper “On the Effect of Populations in Evolutionary Multi-Objective Optimisation” by Oliver Giel and Per Kristian Lehre studies the basic question whether a population is provably useful to compute the whole Pareto front of a multi-objective optimization problem. They compare restart strategies of several single-objective approaches based on one individual to a natural multi-objective approach using a population. Their analyses point out the benefits of the multi-objective evolutionary algorithm in a rigorous way. Using illustrative example functions they provide not only proven results but also a good intuition about why population-based approaches can outperform individual-based approaches in multi-objective optimization. The paper “Exploring the Runtime of an Evolutionary Algorithm for the Multi-Objective Shortest Path Problem” by Christian Horoba examines the behavior of evolutionary algorithms for the classical NP-hard multi-objective shortest path problem. The author considers an evolutionary approach based on -dominance and proves that this algorithm constitutes a fully polynomial-time randomized approximation scheme (FPRAS) for the problem. The algorithm achieves this by mimicking a problemspecific algorithm for that problem. This provides general insight into why evolutionary algorithms can perform well for problems that allow for efficient problem-specific algorithms. Hypervolume-based algorithms have become very popular in evolutionary multiobjective optimization. In the paper, “An Efficient Algorithm for Computing Hypervolume Contributions,” Karl Bringmann and Tobias Friedrich present a faster algorithm for determining the individual with the smallest hypervolume contribution in a given population. Furthermore, they study the problem of removing λ > 1 individuals with the smallest hypervolume contribution from a given population and present an elegant and fast method to carry out this task.
Read moreInvestigation and Comparison of the Performance of Multi-Objective Evolutionary Algorithms Based on Decomposition and Dominance in Portfolio Optimization
A new approach in multi-objective evolutionary optimization is decomposition. Decomposition is a basic method in old multi-objective optimization which in evolutionary multiobjective optimization, has not been widely used. Using this method, a multi-objective optimization problem is converted into a number of scalar subproblems, and all the subproblems are simultaneously optimized. In this paper, the performance and efficiency of the algorithm MOEA/D (multi-objective evolutionary algorithm based on decomposition) with the performance of two algorithms NSGA-II and MOPSO (evolution optimization methods based on dominance) for solving constrained portfolio optimization in Tehran Stock Exchange, has been compared. Portfolio return and its risk, has been considered as the optimization objectives and CvaR has been considered as a risk measure. The results indicate the high potential of these algorithms for constrained portfolio optimization. Also, the results indicate that the optimization algorithm based on decomposition has lower computational complexity, and Pareto front is more extensive than the other two methods. (Abstract)
Read moreEvolutionary diversity optimization using multi-objective indicators
Evolutionary diversity optimization aims to compute a set of solutions that are diverse in the search space or instance feature space, and where all solutions meet a given quality criterion. With this paper, we bridge the areas of evolutionary diversity optimization and evolutionary multi-objective optimization. We show how popular indicators frequently used in the area of multi-objective optimization can be used for evolutionary diversity optimization. Our experimental investigations for evolving diverse sets of TSP instances and images according to various features show that two of the most prominent multi-objective indicators, namely the hypervolume indicator and the inverted generational distance, provide excellent results in terms of visualization and various diversity indicators.
Read moreLearning the Search Range for Evolutionary Optimization in Dynamic Environments
Conventional evolutionary algorithms operate in a fixed search space with limiting parameter range, which is often predefined via a priori knowledge or trial and error in order to `guess' a suitable region comprising the global optimal solution. This requirement is hard, if not impossible, to fulfil in many real-world optimization problems since there is often no clue of where the desired solutions are located in these problems. Thus, this paper proposes an inductive---deductive learning approach for single- and multi-objective evolutionary optimization. The method is capable of directing evolution towards more promising search regions even if these regions are outside the initial predefined space. For problems where the global optimum is included in the initial search space, it is capable of shrinking the search space dynamically for better resolution in genetic representation to facilitate the evolutionary search towards more accurate optimal solutions. Validation results based on benchmark optimization problems show that the proposed inductive---deductive learning is capable of handling different fitness landscapes as well as distributing nondominated solutions uniformly along the final trade-offs in multi-objective optimization, even if there exist many local optima in a high-dimensional search space or the global optimum is outside the predefined search region.
Read moreInteractive evolutionary multiobjective optimization driven by robust ordinal regression
Interactive evolutionary multiobjective optimization driven by robust ordinal regressionThis paper presents the Necessary-preference-enhanced Evolutionary Multiobjective Optimizer (NEMO), which combines an evolutionary multiobjective optimization with robust ordinal regression within an interactive procedure. In the course of NEMO, the decision maker is asked to express preferences by simply comparing some pairs of solutions in the current population. The whole set of additive value functions compatible with this preference information is used within a properly modified version of the evolutionary multiobjective optimization technique NSGA-II in order to focus the search towards solutions satisfying the preferences of the decision maker. This allows to speed up convergence to the most preferred region of the Pareto-front.
Read moreMaintaining Diversity in The Bounded Pareto-Set
For more than two decades, stand-alone evolutionary multi-objective optimization (EMO) methods have been adequately demonstrated to find a set of trade-off solutions near Pareto-front for various multi-objective optimization problems. Despite a long-standing existence of classical generative single-objective based methods, a very few EMO studies have combined the two approaches for a better gain. In this paper, we investigate the effect of seeding the initial population of an EMO algorithm with extreme solutions obtained using a single-objective method. Our proposed approach is further aided with an opposition based offspring creation mechanism which strategically places new solutions on the current Pareto frontier by a simple, yet a novel arbitration policy that utilizes the relative distances from the extreme solutions in the current population members. We conduct an extensive simulation of our proposed approach on a wide variety of two and three-objective benchmark MOP test problems. Results are shown to be remarkably better than the original EMO approach in terms of hyper-volume metric. The results are interesting and should motivate EMO researchers to integrate single-objective focused optimization and an opposition-based concept with diversity-preserving EMO procedures for an overall better performance.
Read moreMultiobjective fuzzy genetics-based machine learning based on MOEA/D with its modifications
Various evolutionary multiobjective optimization (EMO) algorithms have been used in the field of evolutionary fuzzy systems (EFS), because EMO algorithms can easily handle multiple objective functions such as the accuracy maximization and complexity minimization for fuzzy system design. Most EMO algorithms used in EFS are Pareto dominance-based algorithms such as NSGA-II, SPEA2, and PAES. There are a few studies where other types of EMO algorithms are used in EFS. In this paper, we apply a multiobjective evolutionary algorithm based on decomposition called MOEA/D to EFS for fuzzy classifier design. MOEA/D is one of the most well-known decomposition-based EMO algorithms. The key idea is to divide a multiobjective optimization problem into a number of single-objective problems using a set of uniformly distributed weight vectors in a scalarizing function. We propose a new scalarizing function called an accuracy-oriented function (AOF) which is specialized for classifier design. We examine the effects of using AOF in MOEA/D on the search ability of our multiobjective fuzzy genetics-based machine learning (GBML). We also examine the synergy effect of MOEA/D with AOF and parallel distributed implementation of fuzzy GBML on the generalization ability.
Read moreA Spectral Clustering-Based Multi-Source Mating Selection Strategy in Evolutionary Multi-Objective Optimization
In evolutionary multi-objective optimization, it has been illuminated that guide search with neighboring solutions improve the quality of new trial solutions and accelerate algorithms convergence by the regularity property of the continuous multi-objective optimization problems (MOPs). Very recently, clustering learning-based mating strategies have been popular for establishing reproduction operators with neighboring solutions. However, the existing mating strategies may be more reasonable with the full consideration and utilization of the regularity property. The current mating restrictions excessively emphasize algorithm convergence and ignore population diversity. In addition, the selected clustering algorithms in mating restrictions are not conducive for solving MOPs, which have complex Pareto sets (PSs) and\or Pareto fronts (PFs). To solve above problems and address both the algorithm convergence and the population diversity of multi-objective evolutionary algorithms (MOEAs), the spectral clustering based multi-source mating selection strategy (SMMS) is designed to detect regularity properties and balance population diversity while accelerating algorithm convergence. Consequently, a spectral clustering based multi-source mating selection multi-objective evolutionary algorithms is proposed, teamed SMMEA. SMMEA is applied to a number of test suites with a complex PS or PF, and compared with six state-of-the-art MOEAs. The results demonstrate that the proposed algorithm outperforms over the other approaches.
Read moreHiding Structure of Reconstructed Network from Subgraphs Using Evolutionary Multi-Objective Optimization
This study considered the problem of reconstructing an overall graph from multiple partially disclosed subgraphs such as supply chains and social networks. Modifying a part of a subgraph can make grasping the structure of the overall graph a challenge, thereby reducing the possibility of identifying essential nodes. However, there is a trade-off between the accuracy of modified subgraphs and the difficulty of grasping the overall structure. We formulated this as a multi-objective optimization problem and investigated its characteristics and the performance of the search operators using evolutionary multi-objective optimization and synthetic problems.
Read moreBalancing exploration and exploitation with adaptive variation for evolutionary multi-objective optimization
Balancing exploration and exploitation with adaptive variation for evolutionary multi-objective optimization
Population Diversity Dynamics Analysis for Imbalanced Multi-objective Optimization
Within the field of evolutionary multi-objective optimization, there exists a class of problems where most evolutionary multi-objective optimization algorithms (EMOAs) suffer from a loss of diversity that remains unrecoverable for extended periods, namely “imbalanced problems”. Currently, the field of multi-objective optimization lacks rigorous quantitative characterization of imbalanced multi-objective optimization problems (MOPs) and systematic understanding of their impact on algorithmic performance. To address these limitations, we introduce two novel quantities: the global and local shrinkage-spread rates, which characterize population convergence and diversity dynamics during global exploration and local exploitation phases, respectively. Based on these quantities, we provide the first mathematical characterization of imbalanced MOPs and derive key theoretical properties through rigorous probabilistic analysis. Our theoretical result shows that the probability of achieving substantial diversity recovery is below 0.1 under given conditions, demonstrating that diversity recovery becomes extremely challenging once population diversity is lost. We further construct ten imbalanced MOPs in which the difficulty of maintaining population diversity can be precisely modulated through adjustable parameters. Using these benchmark problems, we conduct comparative experiments evaluating four representative EMOAs, i.e., NSGA-II, MOEA/D, RVEA, MOEA/D-M2M and DrEA, in terms of their diversity maintenance capabilities on imbalanced MOPs. The results provide both theoretical insights and practical guidance for designing robust EMOAs capable of handling imbalanced MOPs.
Read moreAssociation Rule Hiding Based on Evolutionary Multi-Objective Optimization by Removing Items
Today, people benefit from utilizing data mining technologies, such as association rule mining methods, to find valuable knowledge residing in a large amount of data. However, they also face the risk of exposing sensitive or confidential information, when data is shared among different organizations. Thus, a question arise: how can we prevent that sensitive knowledge is discovered, while ensuring that ordinary non-sensitive knowledge can be mined to the maximum extent possible. In this paper, we address the problem of privacy preserving in association rule mining from the perspective of multi-objective optimization. A new hiding method based evolutionary multi-objective optimization (EMO) is proposed and the side effects generated by the hiding process are formulated as optimization goals. EMO is used to find candidate transactions to modify so that side effects are minimized. Comparative experiments with exact methods on real datasets demonstrated that the proposed method can hide sensitive rules with fewer side effects.
Read moreModelling the Pareto-optimal set using B-spline basis functions for continuous multi-objective optimization problems
In the past few years, multi-objective optimization algorithms have been extensively applied in several fields including engineering design problems. A major reason is the advancement of evolutionary multi-objective optimization (EMO) algorithms that are able to find a set of non-dominated points spread on the respective Pareto-optimal front in a single simulation. Besides just finding a set of Pareto-optimal solutions, one is often interested in capturing knowledge about the variation of variable values over the Pareto-optimal front. Recent innovization approaches for knowledge discovery from Pareto-optimal solutions remain as a major activity in this direction. In this article, a different data-fitting approach for continuous parameterization of the Pareto-optimal front is presented. Cubic B-spline basis functions are used for fitting the data returned by an EMO procedure in a continuous variable space. No prior knowledge about the order in the data is assumed. An automatic procedure for detecting gaps in the Pareto-optimal front is also implemented. The algorithm takes points returned by the EMO as input and returns the control points of the B-spline manifold representing the Pareto-optimal set. Results for several standard and engineering, bi-objective and tri-objective optimization problems demonstrate the usefulness of the proposed procedure.
Read more