scholarly journals An evolutionary decomposition-based multi-objective feature selection for multi-label classification

2020 ◽  
Vol 6 ◽  
pp. e261
Author(s):  
Azam Asilian Bidgoli ◽  
Hossein Ebrahimpour-Komleh ◽  
Shahryar Rahnamayan
Keyword(s):  
Feature Selection ◽  
Real World ◽  
Classification Accuracy ◽  
Optimization Approach ◽  
Multi Objective Optimization ◽  
Classification Problems ◽  
Genetic Operators ◽  
Feature Selection Problem ◽  
Multi Objective ◽  
Conflicting Objectives

Data classification is a fundamental task in data mining. Within this field, the classification of multi-labeled data has been seriously considered in recent years. In such problems, each data entity can simultaneously belong to several categories. Multi-label classification is important because of many recent real-world applications in which each entity has more than one label. To improve the performance of multi-label classification, feature selection plays an important role. It involves identifying and removing irrelevant and redundant features that unnecessarily increase the dimensions of the search space for the classification problems. However, classification may fail with an extreme decrease in the number of relevant features. Thus, minimizing the number of features and maximizing the classification accuracy are two desirable but conflicting objectives in multi-label feature selection. In this article, we introduce a multi-objective optimization algorithm customized for selecting the features of multi-label data. The proposed algorithm is an enhanced variant of a decomposition-based multi-objective optimization approach, in which the multi-label feature selection problem is divided into single-objective subproblems that can be simultaneously solved using an evolutionary algorithm. This approach leads to accelerating the optimization process and finding more diverse feature subsets. The proposed method benefits from a local search operator to find better solutions for each subproblem. We also define a pool of genetic operators to generate new feature subsets based on old generation. To evaluate the performance of the proposed algorithm, we compare it with two other multi-objective feature selection approaches on eight real-world benchmark datasets that are commonly used for multi-label classification. The reported results of multi-objective method evaluation measures, such as hypervolume indicator and set coverage, illustrate an improvement in the results obtained by the proposed method. Moreover, the proposed method achieved better results in terms of classification accuracy with fewer features compared with state-of-the-art methods.

Multi-Objective Optimization in Multi-Attribute and Multi-Unit Combinatorial Reverse Auctions

2017 ◽  
Vol 26 (05) ◽  
pp. 1760016 ◽  
Author(s):  
Shubhashis Kumar Shil ◽  
Samira Sadaoui
Keyword(s):  
Quality Measures ◽  
Optimization Approach ◽  
Multi Objective Optimization ◽  
Solution Quality ◽  
Winner Determination ◽  
Multi Objective ◽  
Conflicting Objectives ◽  
Trading Constraints ◽  
Significant Superiority

This study introduces an advanced Combinatorial Reverse Auction (CRA), multi-units, multiattributes and multi-objective, which is subject to buyer and seller trading constraints. Conflicting objectives may occur since the buyer can maximize some attributes and minimize some others. To address the Winner Determination (WD) problem for this type of CRAs, we propose an optimization approach based on genetic algorithms that we integrate with our variants of diversity and elitism strategies to improve the solution quality. Moreover, by maximizing the buyer’s revenue, our approach is able to return the best solution for our complex WD problem. We conduct a case study as well as simulated testing to illustrate the importance of the diversity and elitism schemes. We also validate the proposed WD method through simulated experiments by generating large instances of our CRA problem. The experimental results demonstrate on one hand the performance of our WD method in terms of several quality measures, like solution quality, run-time complexity and trade-off between convergence and diversity, and on the other hand, it’s significant superiority to well-known heuristic and exact WD techniques that have been implemented for much simpler CRAs.

scholarly journals Handling Constrained Multi-objective Optimization By Ignoring Constraints and Using Two Evolutionary Frameworks

2020 ◽  
Author(s):  
Xiang Yi ◽  
Xiaowei Yang ◽  
Han Huang ◽  
Jiahai Wang
Keyword(s):  
Real World ◽  
Optimization Problems ◽  
Evolutionary Process ◽  
Inequality Constraints ◽  
Simultaneous Optimization ◽  
Multi Objective Optimization ◽  
Multi Objective ◽  
Real World Problem ◽  
Conflicting Objectives ◽  
Two Stages

Constrained multi-objective optimization problems exist widely in real-world applications, and they involve a simultaneous optimization of multiple and often conflicting objectives subject to several equality and/or inequality constraints. To deal with these problems, a crucial issue is how to handle constraints effectively. This paper proposes a simple yet effective constrained decomposition-based multi-objective evolutionary algorithm. In the proposal, the evolutionary process is divided into two stages in which constraints are handled differently. In the first stage, constraints are totally ignored and the population is pulled toward the unconstrained Pareto-optimal front (PF) by optimizing objectives only. This can help the proposed algorithm handle well problems with the following features, i.e., the constrained PF has an intersection with the unconstrained counterpart, and there are infeasible regions blocking the way of convergence. In the second stage, with the purpose of approximating the constrained PF well,constraint satisfaction is emphasized over objective minimization.Moreover, different evolutionary frameworks are adopted in the two stages to promote the performance of the algorithm as much as possible. The proposed algorithm is comprehensively compared with several state-of-the-art algorithms on 39 problems (with 266 test instances in total), including one real-world problem (with 36 instances) in search-based software engineering. As shown by the experimental results, the new algorithm performs best on the majority of these problems, particularly on those with the aforementioned features. In summary, the suggested algorithm provides an effective way of handling constrained multi-objective optimization problems.

scholarly journals Classification of Reservoir Recovery Factor for Oil and Gas Reservoirs: A Multi-Objective Feature Selection Approach

2021 ◽  
Vol 9 (8) ◽  
pp. 888
Author(s):  
Qasem Al-Tashi ◽  
Emelia Akashah Patah Akhir ◽  
Said Jadid Abdulkadir ◽  
Seyedali Mirjalili ◽  
Tareq M. Shami ◽  
...  
Keyword(s):  
Feature Selection ◽  
Recovery Factor ◽  
Classification Error ◽  
Grey Wolf Optimizer ◽  
Small Subset ◽  
Multi Objective Optimization ◽  
Nsga Ii ◽  
Multi Objective ◽  
Conflicting Objectives

The accurate classification of reservoir recovery factor is dampened by irregularities such as noisy and high-dimensional features associated with the reservoir measurements or characterization. These irregularities, especially a larger number of features, make it difficult to perform accurate classification of reservoir recovery factor, as the generated reservoir features are usually heterogeneous. Consequently, it is imperative to select relevant reservoir features while preserving or amplifying reservoir recovery accuracy. This phenomenon can be treated as a multi-objective optimization problem, since there are two conflicting objectives: minimizing the number of measurements and preserving high recovery classification accuracy. In this study, wrapper-based multi-objective feature selection approaches are proposed to estimate the set of Pareto optimal solutions that represents the optimum trade-off between these two objectives. Specifically, three multi-objective optimization algorithms—Non-dominated Sorting Genetic Algorithm II (NSGA-II), Multi-Objective Grey Wolf Optimizer (MOGWO) and Multi-Objective Particle Swarm Optimization (MOPSO)—are investigated in selecting relevant features from the reservoir dataset. To the best of our knowledge, this is the first time multi-objective optimization has been used for reservoir recovery factor classification. The Artificial Neural Network (ANN) classification algorithm is used to evaluate the selected reservoir features. Findings from the experimental results show that the proposed MOGWO-ANN outperforms the other two approaches (MOPSO and NSGA-II) in terms of producing non-dominated solutions with a small subset of features and reduced classification error rate.

Design of Optimal Operational Parameters for Steam-Alternating-Solvent Processes in Heterogeneous Reservoirs - A Multi-Objective Optimization Approach

2021 ◽  
Author(s):  
Israel Mayo-Molina ◽  
Juliana Y. Leung
Keyword(s):  
Sensitivity Analysis ◽  
Computational Time ◽  
Optimization Approach ◽  
Objective Functions ◽  
Multi Objective Optimization ◽  
Operational Parameters ◽  
Multi Objective ◽  
Bottom Hole ◽  
Decision Variables ◽  
Conflicting Objectives

Abstract The Steam Alternating Solvent (SAS) process has been proposed and studied in recent years as a new auspicious alternative to the conventional thermal (steam-based) bitumen recovery process. The SAS process incorporates steam and solvent (e.g. propane) cycles injected alternatively using the same configuration as the Steam-Assisted Gravity-Drainage (SAGD) process. The SAS process offers many advantages, including lower capital and operational cost, as well as a reduction in water usage and lower Greenhouse Gas (GHG) Emissions. On the other hand, one of the main challenges of this relatively new process is the influence of uncertain reservoir heterogeneity distribution, such as shale barriers, on production behaviour. Many complex physical mechanisms, including heat transfer, fluid flows, and mass transfer, must be coupled. A proper design and selection of the operational parameters must consider several conflicting objectives. This work aims to develop a hybrid multi-objective optimization (MOO) framework for determining a set of Pareto-optimal SAS operational parameters under a variety of heterogeneity scenarios. First, a 2-D homogeneous reservoir model is constructed based on typical Cold lake reservoir properties in Alberta, Canada. The homogeneous model is used to establish a base scenario. Second, different shale barrier configurations with varying proportions, lengths, and locations are incorporated. Third, a detailed sensitivity analysis is performed to determine the most impactful parameters or decision variables. Based on the results of the sensitivity analysis, several objective functions are formulated (e.g., minimizing energy and solvent usage). Fourth, Response Surface Methodology (RSM) is applied to generate a set of proxy models to approximate the non-linear relationship between the decision variables and the objective functions and to reduce the overall computational time. Finally, three Multi-Objective Evolutionary Algorithms (MOEAs) are applied to search and compare the optimal sets of decision parameters. The study showed that the SAS process is sensitive to the shale barrier distribution, and that impact is strongly dependent on the location and length of a specific shale barrier. When a shale barrier is located near the injector well, pressure and temperature may build up in the near-well area, preventing additional steam and solvent be injected and, consequently, reducing the oil production. Operational constraints, such as bottom-hole pressure, steam trap criterion, and bottom-hole gas rate in the producer, are among various critical decision variables examined in this study. A key conclusion is that the optimal operating strategy should depend on the underlying heterogeneity. Although this notion has been alluded to in other previous steam- or solvent-based studies, this paper is the first to utilize a MOO framework for systematically determining a specific optimal strategy for each heterogeneity scenario. With the advancement of continuous downhole fibre-optic monitoring, the outcomes can potentially be integrated into other real-time reservoir characterization and optimization work-flows.

A Multi-Objective Optimization Approach on Spiral Grooves for Gas Mechanical Seals

2018 ◽  
Vol 140 (4) ◽  
Author(s):  
Xiuying Wang ◽  
Liping Shi ◽  
Wei Huang ◽  
Xiaolei Wang
Keyword(s):  
Optimization Approach ◽  
Multi Objective Optimization ◽  
Mechanical Seals ◽  
Independent Variables ◽  
Multi Objective ◽  
Pareto Optimal Set ◽  
Conflicting Objectives ◽  
Collinearity Diagnostics ◽  
Film Stiffness ◽  
Spiral Grooves

Spiral groove is one of the most common types of structures on gas mechanical seals. Numerical research demonstrated that the grooves designed for improving gas film lift or film stiffness often lead to the leakage increase. Hence, a multi-objective optimization approach specially for conflicting objectives is utilized to optimize the spiral grooves for a specific sample in this study. First, the objectives and independent variables in multi-objective optimization are determined by single objective analysis. Then, a set of optimal parameters, i.e., Pareto-optimal set, is obtained. Each solution in this set can get the highest dimensionless gas film lift under a specific requirement of the dimensionless leakage rate. Finally, the collinearity diagnostics is performed to evaluate the importance of different independent variables in the optimization.

Pareto-Based Multi-Objective Optimization of Recuperated S-CO2 Brayton Cycles

2014 ◽  
Author(s):  
Mahmood Mohagheghi ◽  
Jayanta Kapat ◽  
Narasimha Nagaiah
Keyword(s):  
System Optimization ◽  
Superior Performance ◽  
Specific Power ◽  
Pareto Optimal ◽  
Optimization Approach ◽  
Multi Objective Optimization ◽  
Multi Objective ◽  
Conflicting Objectives ◽  
Cycle Efficiency ◽  
Pareto Optimal Fronts

In this paper, two configurations of the S-CO2 Brayton cycles (i.e., the single-recuperated and recompression cycles) are thermodynamically modeled and optimized through a multi-objective approach. Two semi-conflicting objectives, i.e., cycle efficiency (ηc) and cycle specific power (Φsp) are maximized simultaneously to achieve Pareto optimal fronts. The objective of maximum cycle efficiency is to have a smaller and less expensive solar field, and a lower fuel cost in case of a hybrid scheme. On the other hand, the objective of maximum specific power provides a smaller power block, and a lower capital cost associated with recuperators and coolers. The multi-objective optimization is carried out by means of a genetic algorithm which is a robust method for multidimensional, nonlinear system optimization. The optimization process is comprehensive, i.e., all the decision variables including the inlet temperatures and pressures of turbines and compressors, the pinch point temperature differences, and the mass flow fraction of the main compressor are optimized simultaneously. The presented Pareto optimal fronts provide two optimum trade-off curves enabling decision makers to choose their desired compromise between the objectives, and to avoid naive solution points obtained from a single-objective optimization approach. Moreover, the comparison of the Pareto optimal fronts associated with the studied configurations reveals the optimum operational region of the recompression configuration where it presents superior performance over the single-recuperated cycle.

MULTI-OBJECTIVE EVOLUTIONARY ALGORITHMS FOR FILTER BASED FEATURE SELECTION IN CLASSIFICATION

2013 ◽  
Vol 22 (04) ◽  
pp. 1350024 ◽  
Author(s):  
BING XUE ◽  
LIAM CERVANTE ◽  
LIN SHANG ◽  
WILL N. BROWNE ◽  
MENGJIE ZHANG
Keyword(s):  
Feature Selection ◽  
Evaluation Criteria ◽  
Classification Performance ◽  
Classification Problems ◽  
Multi Objective ◽  
Conflicting Objectives ◽  
Benchmark Datasets ◽  
Selection Algorithms ◽  
Filter Algorithms ◽  
Single Objective

Feature selection is a multi-objective problem with the two main conflicting objectives of minimising the number of features and maximising the classification performance. However, most existing feature selection algorithms are single objective and do not appropriately reflect the actual need. There are a small number of multi-objective feature selection algorithms, which are wrapper based and accordingly are computationally expensive and less general than filter algorithms. Evolutionary computation techniques are particularly suitable for multi-objective optimisation because they use a population of candidate solutions and are able to find multiple non-dominated solutions in a single run. However, the two well-known evolutionary multi-objective algorithms, non-dominated sorting based multi-objective genetic algorithm II (NSGAII) and strength Pareto evolutionary algorithm 2 (SPEA2) have not been applied to filter based feature selection. In this work, based on NSGAII and SPEA2, we develop two multi-objective, filter based feature selection frameworks. Four multi-objective feature selection methods are then developed by applying mutual information and entropy as two different filter evaluation criteria in each of the two proposed frameworks. The proposed multi-objective algorithms are examined and compared with a single objective method and three traditional methods (two filters and one wrapper) on eight benchmark datasets. A decision tree is employed to test the classification performance. Experimental results show that the proposed multi-objective algorithms can automatically evolve a set of non-dominated solutions that include a smaller number of features and achieve better classification performance than using all features. NSGAII and SPEA2 outperform the single objective algorithm, the two traditional filter algorithms and even the traditional wrapper algorithm in terms of both the number of features and the classification performance in most cases. NSGAII achieves similar performance to SPEA2 for the datasets that consist of a small number of features and slightly better results when the number of features is large. This work represents the first study on NSGAII and SPEA2 for filter feature selection in classification problems with both providing field leading classification performance.

scholarly journals A multi-objective artificial bee colony approach to feature selection using fuzzy mutual information

2021 ◽  
Author(s):  
E Hancer ◽  
Bing Xue ◽  
Mengjie Zhang ◽  
D Karaboga ◽  
B Akay
Keyword(s):  
Feature Selection ◽  
Mutual Information ◽  
Classification Accuracy ◽  
Artificial Bee Colony ◽  
Feature Subset ◽  
Fitness Evaluation ◽  
Multi Objective ◽  
Bee Colony ◽  
Conflicting Objectives ◽  
Single Objective

© 2015 IEEE. Feature selection often involves two conflicting objectives of minimizing the feature subset size and the maximizing the classification accuracy. In this paper, a multi-objective artificial bee colony (MOABC) framework is developed for feature selection in classification, and a new fuzzy mutual information based criterion is proposed to evaluate the relevance of feature subsets. Three new multi-objective feature selection approaches are proposed by integrating MOABC with three filter fitness evaluation criteria, which are mutual information, fuzzy mutual information and the proposed fuzzy mutual information. The proposed multi-objective feature selection approaches are examined by comparing them with three single-objective ABC-based feature selection approaches on six commonly used datasets. The results show that the proposed approaches are able to achieve better performance than the original feature set in terms of the classification accuracy and the number of features. By using the same evaluation criterion, the proposed multi-objective algorithms generally perform better than the single-objective methods, especially in terms of reducing the number of features. Furthermore, the proposed fuzzy mutual information criterion outperforms mutual information and the original fuzzy mutual information in both single-objective and multi-objective manners. This work is the first study on multi-objective ABC for filter feature selection in classification, which shows that multi-objective ABC can be effectively used to address feature selection problems.

Handling Constrained Multi-objective Optimization By Ignoring Constraints and Using Two Evolutionary Frameworks

2020 ◽  
Author(s):  
Xiang Yi ◽  
Xiaowei Yang ◽  
Han Huang ◽  
Jiahai Wang
Keyword(s):  
Real World ◽  
Optimization Problems ◽  
Evolutionary Process ◽  
Inequality Constraints ◽  
Simultaneous Optimization ◽  
Multi Objective Optimization ◽  
Multi Objective ◽  
Real World Problem ◽  
Conflicting Objectives ◽  
Two Stages

Constrained multi-objective optimization problems exist widely in real-world applications, and they involve a simultaneous optimization of multiple and often conflicting objectives subject to several equality and/or inequality constraints. To deal with these problems, a crucial issue is how to handle constraints effectively. This paper proposes a simple yet effective constrained decomposition-based multi-objective evolutionary algorithm. In the proposal, the evolutionary process is divided into two stages in which constraints are handled differently. In the first stage, constraints are totally ignored and the population is pulled toward the unconstrained Pareto-optimal front (PF) by optimizing objectives only. This can help the proposed algorithm handle well problems with the following features, i.e., the constrained PF has an intersection with the unconstrained counterpart, and there are infeasible regions blocking the way of convergence. In the second stage, with the purpose of approximating the constrained PF well,constraint satisfaction is emphasized over objective minimization.Moreover, different evolutionary frameworks are adopted in the two stages to promote the performance of the algorithm as much as possible. The proposed algorithm is comprehensively compared with several state-of-the-art algorithms on 39 problems (with 266 test instances in total), including one real-world problem (with 36 instances) in search-based software engineering. As shown by the experimental results, the new algorithm performs best on the majority of these problems, particularly on those with the aforementioned features. In summary, the suggested algorithm provides an effective way of handling constrained multi-objective optimization problems.

Sign in / Sign up

Export Citation Format

Share Document