Complex problems can be analyzed by using model simulation but its use is not straight-forward since modelers must carefully calibrate and validate their models before using them. This is specially relevant for models...
详细信息
Complex problems can be analyzed by using model simulation but its use is not straight-forward since modelers must carefully calibrate and validate their models before using them. This is specially relevant for models considering multiple outputs as its calibration requires handling different criteria jointly. This can be achieved using automated calibration and evolutionary multiobjective optimization methods which are the state of the art in multiobjectiveoptimization as they can find a set of representative Pareto solutions under these restrictions and in a single run. However, selecting the best algorithm for performing automated calibration can be overwhelming. We propose to deal with this issue by conducting an exhaustive analysis of the performance of several evolutionary multiobjective optimization algorithms when calibrating several instances of an agent-based model for marketing with multiple outputs. We analyze the calibration results using multiobjective performance indicators and attainment surfaces, including a statistical test for studying the significance of the indicator values, and benchmarking their performance with respect to a classical mathematical method. The results of our experimentation reflect that those algorithms based on decomposition perform significantly better than the remaining methods in most instances. Besides, we also identify how different properties of the problem instances (i.e., the shape of the feasible region, the shape of the Pareto front, and the increased dimensionality) erode the behavior of the algorithms to different degrees.
evolutionary multiobjective optimization has witnessed remarkable progress during the past decades. However, existing algorithms often encounter computational challenges in large-scale scenarios, primarily attributed ...
详细信息
ISBN:
(纸本)9798400704949
evolutionary multiobjective optimization has witnessed remarkable progress during the past decades. However, existing algorithms often encounter computational challenges in large-scale scenarios, primarily attributed to the absence of hardware acceleration. In response, we introduce a Tensorized Reference Vector Guided evolutionary Algorithm (TensorRVEA) for harnessing the advancements of GPU acceleration. In TensorRVEA, the key data structures and operators are fully transformed into tensor forms for leveraging GPU-based parallel computing. In numerical benchmark tests involving large-scale populations and problem dimensions, TensorRVEA consistently demonstrates high computational performance, achieving up to over 1000x speedups. Then, we applied TensorRVEA to the domain of multiobjective neuroevolution for addressing complex challenges in robotic control tasks. Furthermore, we assessed TensorRVEA's extensibility by altering several tensorized reproduction operators. Experimental results demonstrate promising scalability and robustness of TensorRVEA. Source codes are available at https://***/EMI-Group/tensorrvea.
The choice of primer designs for polymerase chain reaction experiments affects the results. Designing optimal combinations of forward and reverse primers requires solving multiple conflicting objectives simultaneously...
详细信息
ISBN:
(纸本)9798400704949
The choice of primer designs for polymerase chain reaction experiments affects the results. Designing optimal combinations of forward and reverse primers requires solving multiple conflicting objectives simultaneously. Most of the tools for primer design optimize the problem by a priori scalarization or by setting constraints with preset preferences. Therefore, the decision-maker (DM) or domain expert has to re-execute the optimizer with new preferences to find satisfactory solutions. An a priori method is detrimental to decision-making since the DM cannot learn about the problem characteristics, and re-executing the optimizer with new preferences increases the number of function evaluations. In addition, the existing methods rely on a single mathematical model to estimate the melting temperature of primers. In this paper, we formulate a multiobjectiveoptimization problem consisting of three uncertain objectives that use six different models to estimate the melting temperatures of primers. The formulated problem was solved using an interactive multiobjectiveevolutionary algorithm that enabled the DM to guide the solution process. We also proposed a selection criterion tailored to our problem that could find optimal primer designs according to the DM's preferences. Finally, we demonstrate the proposed interactive approach to find optimal primers for a bacterial 16S DNA sequence.
Preference-based evolutionary multiobjective optimization (EMO) algorithms approximate the region of interest (ROI) of the Pareto optimal front defined by the preferences of a decision maker (DM). Here, we propose a p...
详细信息
Preference-based evolutionary multiobjective optimization (EMO) algorithms approximate the region of interest (ROI) of the Pareto optimal front defined by the preferences of a decision maker (DM). Here, we propose a preference-based EMO algorithm, in which the preferences are given by means of aspiration and reservation points. The aspiration point is formed by objective values which the DM wants to achieve, while the reservation point is constituted by values for the objectives not to be worsened. Internally, the first generations are performed in order to generate an initial approximation set according to the reservation point. Next, in the remaining generations, the algorithm adapts the search for new non-dominated solutions depending on the dominance relation between the solutions obtained so far and both the reservation and aspiration points. This allows knowing if the given points are achievable or not;this type of information cannot be known before the solution process starts. On this basis, the algorithm proceeds according to three different scenarios with the aim of re-orienting the search directions towards the ROI formed by the Pareto optimal solutions with objective values within the given aspiration and reservation values. Computational results show the potential of our proposal in 2, 3 and 5-objective test problems, in comparison to other state-of-the-art algorithms.
It is particularly challenging for evolutionary algorithms to quickly converge to the Pareto front in large-scale multiobjectiveoptimization. To tackle this problem, this article proposes a large-scale multiobjective...
详细信息
It is particularly challenging for evolutionary algorithms to quickly converge to the Pareto front in large-scale multiobjectiveoptimization. To tackle this problem, this article proposes a large-scale multiobjectiveevolutionary algorithm assisted by some selected individuals generated by directed sampling (DS). At each generation, a set of individuals closer to the ideal point is chosen for performing a DS in the decision space, and those nondominated ones of the sampled solutions are used to assist the reproduction to improve the convergence in evolutionary large-scale multiobjectiveoptimization. In addition, elitist nondominated sorting is adopted complementarily for environmental selection with a reference vector-based method in order to maintain diversity of the population. Our experimental results show that the proposed algorithm is highly competitive on large-scale multiobjectiveoptimization test problems with up to 5000 decision variables compared to five state-of-the-art multiobjectiveevolutionary algorithms.
In this paper, we propose a preference-based evolutionary multiobjective optimization algorithm, at which the preferences are given in the form of desirable ranges for the objective functions, i.e. by means of aspirat...
详细信息
ISBN:
(纸本)9783030726980;9783030726997
In this paper, we propose a preference-based evolutionary multiobjective optimization algorithm, at which the preferences are given in the form of desirable ranges for the objective functions, i.e. by means of aspiration and reservation levels. The aspiration levels are values to be achieved by the objectives, while the reservation levels are objective values not to be worsen. In the algorithm proposed, the first generations are performed using a set of weight vectors to initially converge to the region of the Pareto optimal front associated with the point formed with the reservation levels. At a certain moment, these weights are updated using the nondominated solutions generated so far, to redirect the search towards the region which contains the Pareto optimal solutions with objective values among the desirable ranges. To this aim, the remaining number of generations are run using the updated weight vectors and the point formed with the aspiration levels. The computational experiment show the potential of our proposal in 2, 3 and 5-objective problems, in comparison to other state-of-the-art algorithms.
Preselection is an important strategy to improve evolutionary algorithms' performance by filtering out unpromising solutions before fitness evaluations. This paper introduces a preselection strategy based on an ap...
详细信息
ISBN:
(纸本)9781728183923
Preselection is an important strategy to improve evolutionary algorithms' performance by filtering out unpromising solutions before fitness evaluations. This paper introduces a preselection strategy based on an approximated Pareto domination relationship for multiobjectiveevolutionaryoptimization. For each objective, a binary relation between each pair of solutions is constructed based on the current population, and a binary classifier is built based on the binary relation pairs. In this way, an approximated Pareto domination relationship can be defined. When new trial solutions are generated, the approximated Pareto domination is used to select promising solutions, which shall be evaluated by the real objective functions. The new preselection is integrated into two algorithms. The experimental results on two benchmark test suites suggest that the algorithms with preselection outperform their original ones.
In most of the previous works, changed and unchanged regions are detected by analyzing the changes of backscattering coefficients for SAR images, which is termed as binary change detection. In fact, due to the increas...
详细信息
In most of the previous works, changed and unchanged regions are detected by analyzing the changes of backscattering coefficients for SAR images, which is termed as binary change detection. In fact, due to the increase and decrease of backscattering coefficients, the changed regions can be further analyzed as two kinds of changes, which is termed as ternary change detection. In this paper, a change detection method based on evolutionary multiobjective optimization is proposed to automatically perform binary and ternary change detection of multitemporal SAR images. First, the log-likelihood function of the Gaussian mixture model and the Bhattacharyya distance are designed as two objectives, respectively. In particular, a novel measurement method based on Bhattacharyya distance is designed for the ternary change detection task. Not only the separability between each two classes is maximized, but also the Bhattacharyya distance between two changed classes and unchanged class is kept closer to obtain a more balanced classification performance. Then a multiobjectiveoptimization method based on non-dominated sorting is used to optimize these two objectives simultaneously. In the proposed approach, chromosome ranking and perturbation probability selection operators are designed to make high-quality solutions with a high probability of being exploited and improve the performance of the algorithm. In addition, a one-step local search strategy based on the expectation-maximization method is integrated into the proposed algorithm to accelerate the convergence. Experimental results on simulated and real-world datasets demonstrate the effectiveness and robustness of the proposed algorithm.(c) 2022 Elsevier B.V. All rights reserved.
Marketers have an important asset if they effectively target social networks' influentials. They can advertise products or services with free items or discounts to spread positive opinions to other consumers (i.e....
详细信息
Marketers have an important asset if they effectively target social networks' influentials. They can advertise products or services with free items or discounts to spread positive opinions to other consumers (i.e., word-of-mouth). However, main research on choosing the best influentials to target is single-objective and mainly focused on maximizing sales revenue. In this paper we propose a multiobjective approach to the influence maximization problem with the aim of increasing the revenue of viral marketing campaigns while reducing the costs. By using local social network metrics to locate influentials, we apply two evolutionary multiobjective optimization algorithms, NSGA-II and MOEA/D, a multiobjective adaptation of a single-objective genetic algorithm, and a greedy algorithm. Our proposal uses a realistic agent-based market framework to evaluate the fitness of the chromosomes by simulating the viral campaigns. The framework also generates, in a single run, a set of non-dominated solutions that allows marketers to consider multiple targeting options . The algorithms are evaluated on five network topologies and a real data-generated social network, showing that both MOEA/D and NSGA-II outperform the single-objective and the greedy approaches. More interestingly, we show a clear correlation between the algorithms' performance and the diffusion features of the social networks. (C) 2020 Elsevier Ltd. All rights reserved.
The prior knowledge from the problem property can boost the evolutionarymultiobjective opti-mization (EMO). The existing machine learning model for knowledge mining in the EMO has led to enhanced performance on multi...
详细信息
The prior knowledge from the problem property can boost the evolutionarymultiobjective opti-mization (EMO). The existing machine learning model for knowledge mining in the EMO has led to enhanced performance on multiobjectiveoptimization problems with complicated Pareto fronts (MOPs-cPF), but the high computational cost resulting from model training should not be taken as a natural expense. To overcome this drawback, this paper proposes an incremental learning-inspired mating restriction strategy (ILMR) for solving MOP-cPF efficiently. In ILMR, a mating restriction is implemented based on an incremental learning model that establishes the mating pool for each solution in the population and incrementally updates as the population evolves. Specifically, it consists of two interdependent parts, i.e., a learning module and a forgetting module. In one evolutionary cycle, the learning module is used to learn new knowledge from the high-quality offspring solutions, while the forgetting module is utilized to remove the information provided by relatively poor solutions in the population. Moreover, a multiobjectiveevolutionary algorithm with ILMR, named MEILM, is proposed and compared with six state-of-the-art algorithms on a variety of MOP-cPF. The experimental results show that there are significant improvements benefitting from the proposed mating restriction strategy.(c) 2022 Elsevier B.V. All rights reserved.
暂无评论