Evolutionary Computation Framework Research Articles

An Evolutionary Computation Framework for Task Off-and-Downloading Scheduling in Mobile Edge Computing

An Evolutionary Computation Framework for Task Off-and-Downloading Scheduling in Mobile Edge Computing

Contrastive variational auto-encoder driven convergence guidance in evolutionary multitasking

Knowledge transfer is at the core of the Evolutionary Multitasking (EMT) problem, as it exploits the interaction of inter-task common knowledge to accelerate task convergence. However, existing research on EMT is generally based solely on the framework of evolutionary computation, with limited integration of deep learning models. Additionally, few studies have found that utilizing deep learning models to generate individuals for transfer and guide the evolutionary trajectory may promote better convergence of algorithms. To address this research gap, we introduce the MFEA-VC (Multifactorial Evolutionary-Variational Auto-Encoder and Contrastive Learning) algorithm. Individuals are categorized based on task-label and inputted into a VAE, with sampling along feature dimensions. The VAE effectively guides the population towards better search areas by learning the latent trends of the distribution and generating transferred individuals. Simultaneously, a new training objective based on contrastive learning is proposed. This objective regulates the similarity between individuals from the same and different task-label, finely controlling individual features in the latent space. This approach makes the generated superior individuals more interpretable. To verify the superiority of MFEA-VC, we conduct comprehensive empirical studies on multi-task single-objective scenarios. We also validate the effectiveness of our improved loss function through theoretical analysis. The results demonstrate that, compared to state-of-the-art multifactorial algorithms, our method significantly enhances the global search capability during the early evolution stages, achieves excellent convergence results, and exhibits strong adaptability to heterogeneous tasks.

Multi-objective reinforcement learning based on nonlinear scalarization and long-short-term optimization

PurposeMany practical control problems require achieving multiple objectives, and these objectives often conflict with each other. The existing multi-objective evolutionary reinforcement learning algorithms cannot achieve good search results when solving such problems. It is necessary to design a new multi-objective evolutionary reinforcement learning algorithm with a stronger searchability.Design/methodology/approachThe multi-objective reinforcement learning algorithm proposed in this paper is based on the evolutionary computation framework. In each generation, this study uses the long-short-term selection method to select parent policies. The long-term selection is based on the improvement of policy along the predefined optimization direction in the previous generation. The short-term selection uses a prediction model to predict the optimization direction that may have the greatest improvement on overall population performance. In the evolutionary stage, the penalty-based nonlinear scalarization method is used to scalarize the multi-dimensional advantage functions, and the nonlinear multi-objective policy gradient is designed to optimize the parent policies along the predefined directions.FindingsThe penalty-based nonlinear scalarization method can force policies to improve along the predefined optimization directions. The long-short-term optimization method can alleviate the exploration-exploitation problem, enabling the algorithm to explore unknown regions while ensuring that potential policies are fully optimized. The combination of these designs can effectively improve the performance of the final population.Originality/valueA multi-objective evolutionary reinforcement learning algorithm with stronger searchability has been proposed. This algorithm can find a Pareto policy set with better convergence, diversity and density.

EvoX: A Distributed GPU-Accelerated Framework for Scalable Evolutionary Computation

EvoX: A Distributed GPU-Accelerated Framework for Scalable Evolutionary Computation

A multi-granular general evolutionary computation framework by fully utilizing the eliminated particles

A multi-granular general evolutionary computation framework by fully utilizing the eliminated particles

Evolutionary generative design of supercritical airfoils: an automated approach driven by small data

Supercritical airfoils are critical components in the design of commercial wide-body aircraft wings due to their ability to enhance aerodynamic performance in transonic flow regimes. However, traditional design methods for supercritical airfoils can be time-consuming and require significant manual effort, not to mention the high cost associated with computational fluid dynamics analysis. To address these challenges, this paper introduces a highly automated approach for supercritical airfoil design, called Evolutionary Generative Design (EvoGD). The EvoGD approach is based on the framework of Evolutionary Computation and employs a series of sophisticated data-driven generative models incorporated with physical information to iteratively refine initial airfoil shapes, resulting in improved aerodynamic performances and reduced constraint violations. Moreover, to speed up the evaluation of the generated airfoils, a series of accurate and efficient data-driven predictors are utilized. The efficacy of the EvoGD approach was demonstrated through experiments on a dataset of 501 supercritical airfoils, including one baseline design and 500 randomly perturbed airfoils. On average, the generated airfoils showed improved performance in terms of buffet lift coefficient, cruise lift-to-drag ratio, and thickness by 5%, 4%, and 1%, respectively. The best generated airfoil outperformed the baseline design in terms of critical buffet lift coefficient and cruise lift-to-drag ratio by 7.1% and 6.4%, respectively. The entire design process was completed in less than an hour on a personal computer, highlighting the high efficiency and scalability of the EvoGD approach.

Gene regulatory networks with binary weights

An evolutionary computation framework to learn binary threshold networks is presented. Inspired by the recent trend of binary neural networks, where weights and activation thresholds are represented using 1 and -1 such that they can be stored in 1-bit instead of full precision, we explore this approach for gene regulatory network modeling. We test our method by inferring binary threshold networks of two regulatory network models: Quorum sensing systems in bacterium Paraburkholderia phytofirmans PsJN and the fission yeast cell-cycle. We considered differential evolution and particle swarm optimization for the simulations. Results for weights having only 1 and -1 values, and different activation thresholds are presented. Full binary threshold networks were found with minimum error (2 bits), whereas when the binary restriction is relaxed for the activation thresholds, networks with 0 bit error were found.

Machine learning-enabled globally guaranteed evolutionary computation

Evolutionary computation, for example, particle swarm optimization, has impressive achievements in solving complex problems in science and industry; however, an important open problem in evolutionary computation is that there is no theoretical guarantee of reaching the global optimum and general reliability; this is due to the lack of a unified representation of diverse problem structures and a generic mechanism by which to avoid local optima. This unresolved challenge impairs trust in the applicability of evolutionary computation to a variety of problems. Here we report an evolutionary computation framework aided by machine learning, named EVOLER, which enables the theoretically guaranteed global optimization of a range of complex non-convex problems. This is achieved by: (1) learning a low-rank representation of a problem with limited samples, which helps to identify an attention subspace; and (2) exploring this small attention subspace via the evolutionary computation method, which helps to reliably avoid local optima. As validated on 20 challenging benchmarks, this method finds the global optimum with a probability approaching 1. We use EVOLER to tackle two important problems: power grid dispatch and the inverse design of nanophotonics devices. The method consistently reached optimal results that were challenging to achieve with previous state-of-the-art methods. EVOLER takes a leap forwards in globally guaranteed evolutionary computation, overcoming the uncertainty of data-driven black-box methods, and offering broad prospects for tackling complex real-world problems.

The unscented genetic algorithm for fast solution of GA-hard optimization problems

This work introduces the Unscented Genetic Algorithm (U-GA), which combines ideas from evolutionary computation and Kalman filters to devise a novel approach to solve GA-hard problems. The approach is justified based on how other Bayesian methods make strong assumptions on data, which could limit their performance in the long run. U-GA applies theory from unscented Kalman filters to relax this assumptions via Monte-Carlo simulation. The algorithm is explained in detail, showing how unscented Kalman filters equations could be adapted for the evolutionary computation framework. In the experiments, the proposed approach is compared to Bayesian optimization algorithm (BOA) and genetic algorithms (GAs) to investigate the strengths and limitations of U-GA. The results show how U-GA attains better performance than the benchmarks, even when the problem size is increased. Also U-GA attained a considerable speed-up (around 400%) when compared with similar methods.

Adaptive multiple selection strategy for multi-objective particle swarm optimization

Multiple-swarm approach is a quite successful evolutionary computation framework for multi-objective particle swarm optimization algorithm (MOPSO) to solve multi-objective optimization problems (MOPs). However, the main challenge of using this framework lies in the lack of leader selection, resulting in the optimal solutions being distributed loosely, as well as far away from the true Pareto-optimal front. To overcome this problem, a multi-swarm MOPSO with an adaptive multiple selection strategy (MOPSO-AMS) is investigated in this paper. This proposed MOPSO-AMS is able to guide each swarm with a suitable leader to improve the evolutionary performance. The novelties and advantages of MOPSO-AMS include the following three aspects. First, a hierarchical evolutionary state detection mechanism, based on the distribution and dominance information of non-dominated solutions, is designed to obtain the evolutionary state of current iteration. Then, the requirements of evolutionary process can be detected. Second, an adaptive multiple selection strategy, using the evolutionary state information and spatial features of candidate solutions, is developed to select leaders of sub-swarms in multiple evolutionary states. Then, suitable leaders can be selected to keep the balance between convergence and diversity. Third, an adaptive parameter adjustment mechanism, based on the dominance relationship of each particle, is introduced to further improve the evolutionary performance of MOPSO-AMS. Finally, numerical simulations and a practical application are used to validate the analytical results and demonstrate the significant improvement of MOPSO-AMS.

GRAPE: Grammatical Algorithms in Python for Evolution

GRAPE is an implementation of Grammatical Evolution (GE) in DEAP, an Evolutionary Computation framework in Python, which consists of the necessary classes and functions to evolve a population of grammar-based solutions, while reporting essential measures. This tool was developed at the Bio-computing and Developmental Systems (BDS) Research Group, the birthplace of GE, as an easy to use (compared to the canonical C++ implementation, libGE) tool that inherits all the advantages of DEAP, such as selection methods, parallelism and multiple search techniques, all of which can be used with GRAPE. In this paper, we address some problems to exemplify the use of GRAPE and to perform a comparison with PonyGE2, an existing implementation of GE in Python. The results show that GRAPE has a similar performance, but is able to avail of all the extra facilities and functionality found in the DEAP framework. We further show that GRAPE enables GE to be applied to systems identification problems and we demonstrate this on two benchmark problems.

Multi-Objective Software Effort Estimation: A Replication Study

Replication studies increase our confidence in previous results when the findings are similar each time, and help mature our knowledge by addressing both internal and external validity aspects. However, these studies are still rare in certain software engineering fields. In this paper, we replicate and extend a previous study, which denotes the current state-of-the-art for multi-objective software effort estimation, namely CoGEE. We investigate the original research questions with an independent implementation and the inclusion of a more robust baseline (LP4EE), carried out by the first author, who was not involved in the original study. Through this replication, we strengthen both the internal and external validity of the original study. We also answer two new research questions investigating the effectiveness of CoGEE by using four additional evolutionary algorithms (i.e., IBEA, MOCell, NSGA-III, SPEA2) and a well-known Java framework for evolutionary computation, namely JMetal (rather than the previously used R software), which allows us to strengthen the external validity of the original study. The results of our replication confirm that: (1) CoGEE outperforms both baseline and state-of-the-art benchmarks statistically significantly ( <inline-formula><tex-math notation="LaTeX">$p &lt;0.001$</tex-math></inline-formula> ); (2) CoGEE’s multi-objective nature makes it able to reach such a good performance; (3) CoGEE’s estimation errors lie within claimed industrial human-expert-based thresholds. Moreover, our new results show that the effectiveness of CoGEE is generally not limited to nor dependent on the choice of the multi-objective algorithm. Using CoGEE with either NSGA-II, NSGA-III, or MOCell produces human competitive results in less than a minute. The Java version of CoGEE has decreased the running time by over 99.8 percent with respect to its R counterpart. We have made publicly available the Java code of CoGEE to ease its adoption, as well as, the data used in this study in order to allow for future replication and extension of our work.

Matrix-Based Evolutionary Computation

Computational intelligence (CI), including artificial neural network, fuzzy logic, and evolutionary computation (EC), has rapidly developed nowadays. Especially, EC is a kind of algorithm for knowledge creation and problem solving, playing a significant role in CI and artificial intelligence (AI). However, traditional EC algorithms have faced great challenge of heavy computational burden and long running time in large-scale (e.g., with many variables) problems. How to efficiently extend EC algorithms to solve complex problems has become one of the most significant research topics in CI and AI communities. To this aim, this paper proposes a matrix-based EC (MEC) framework to extend traditional EC algorithms for efficiently solving large-scale or super large-scale optimization problems. The proposed framework is an entirely new perspective on EC algorithm, from the solution representation to the evolutionary operators. In this framework, the whole population (containing a set of individuals) is defined as a matrix, where a row stands for an individual and a column stands for a dimension (decision variable). This way, the parallel computing functionalities of matrix can be directly and easily carried out on the high performance computing resources to accelerate the computational speed of evolutionary operators. This paper gives two typical examples of MEC algorithms, named matrix-based genetic algorithm and matrix-based particle swarm optimization. Their matrix-based solution representations are presented and the evolutionary operators based on the matrix are described. Moreover, the time complexity is analyzed and the experiments are conducted to show that these MEC algorithms are efficient in reducing the computational time on large scale of decision variables. The MEC is a promising way to extend EC to complex optimization problems in big data environment, leading to a new research direction in CI and AI.

Evidential evolving C-means clustering method based on artificial bee colony algorithm with variable strings and interactive evaluation mode

The Evidential C-Means algorithm provides a global treatment of ambiguity and uncertainty in memberships when partitioning attribute data, but still requires the number of clusters to be fixed as a priori, like most existing clustering methods do. However, the users usually do not know the exact number of clusters in advance, particularly in practical engineering. To relax this requirement, this paper proposes an Evidential Evolving C-Means (E2CM) clustering method in the framework of evolutionary computation: cluster centers are encoded in a population of variable strings (or particles) to search the optimal number and locations of clusters simultaneously. To perform such joint optimization problem well, an artificial bee colony algorithm with variable strings and interactive evaluation mode is proposed. It will be shown that the E2CM can automatically create appropriate credal partitions by just requiring an upper bound of the cluster number rather than the exact one. More interestingly, there are no restrictions on this upper bound from the theoretic point of view. Some numerical experiments and a practical application in thermal power engineering validate our conclusions.

A Probabilistic Niching Evolutionary Computation Framework Based on Binary Space Partitioning

Multimodal optimization problems have multiple satisfactory solutions to identify. Most of the existing works conduct the search based on the information of the current population, which can be inefficient. This article proposes a probabilistic niching evolutionary computation framework that guides the future search based on more sufficient historical information, in order to locate diverse and high-quality solutions. A binary space partition tree is built to structurally organize the space visiting information. Based on the tree, a probabilistic niching strategy is defined to reinforce exploration and exploitation by making full use of the structural historical information. The proposed framework is universal for incorporating various baseline niching algorithms. In this article, we integrate the proposed framework with two niching algorithms: 1) a distance-based differential evolution algorithm and 2) a topology-based particle swarm optimization algorithm. The two new algorithms are evaluated on 20 multimodal optimization test functions. The experimental results show that the proposed framework helps the algorithms obtain competitive performance. They outperform a number of state-of-the-art niching algorithms on most of the test functions.

Curbing Negative Influences Online for Seamless Transfer Evolutionary Optimization.

This paper draws motivation from the remarkable ability of humans to extract useful building-blocks of knowledge from past experiences and spontaneously reuse them for new and more challenging tasks. It is contended that successfully replicating such capabilities in computational solvers, particularly global black-box optimizers, can lead to significant performance enhancements over the current state-of-the-art. The main challenge to overcome is that in general black-box settings, no problem-specific data may be available prior to the onset of the search, thereby limiting the possibility of offline measurement of the synergy between problems. In light of the above, this paper introduces a novel evolutionary computation framework that enables online learning and exploitation of similarities across optimization problems, with the goal of achieving an algorithmic realization of the transfer optimization paradigm. One of the salient features of our proposal is that it accounts for latent similarities which while being less apparent on the surface, may be gradually revealed during the course of the evolutionary search. A theoretical analysis of our proposed framework is carried out, substantiating its positive influences on optimization performance. Furthermore, the practical efficacy of an instantiation of an adaptive transfer evolutionary algorithm is demonstrated on a series of numerical examples, spanning discrete, continuous, as well as single- and multi-objective optimization.

Multifactorial Evolutionary Algorithm With Online Transfer Parameter Estimation: MFEA-II

Humans rarely tackle every problem from scratch. Given this observation, the motivation for this paper is to improve optimization performance through adaptive knowledge transfer across related problems. The scope for spontaneous transfers under the simultaneous occurrence of multiple problems unveils the benefits of multitasking. Multitask optimization has recently demonstrated competence in solving multiple (related) optimization tasks concurrently. Notably, in the presence of underlying relationships between problems, the transfer of high-quality solutions across them has shown to facilitate superior performance characteristics. However, in the absence of any prior knowledge about the intertask synergies (as is often the case with general black-box optimization), the threat of predominantly negative transfer prevails. Susceptibility to negative intertask interactions can impede the overall convergence behavior. To allay such fears, in this paper, we propose a novel evolutionary computation framework that enables online learning and exploitation of the similarities (and discrepancies) between distinct tasks in multitask settings, for an enhanced optimization process. Our proposal is based on the principled theoretical arguments that seek to minimize the tendency of harmful interactions between tasks, based on a purely data-driven learning of relationships among them. The efficacy of our proposed method is validated experimentally on a series of synthetic benchmarks, as well as a practical study that provides insights into the behavior of the method in the face of several tasks occurring at once.

Multiple Parameters Determination for Image and Video Reconstruction Using Genetic Algorithm and Generalized Stein’s Unbiased Risk Estimation

This paper proposes a heuristic multi-parameter optimization method for image and video reconstruction. The method is embedded in an evolutionary computation framework in which a set of parameters are encoded into chromosomes of an individual, and a generalized Stein’s unbiased risk estimation (GSURE) is employed as a fitness function. By using the proposed method, the near-optimal parameters can be determined after several iterations, and with the use of these parameters, the ideal image can then be reconstructed. Furthermore, this method can be extended to video reconstruction. According to the similarity between adjacent frames of a video, the generation of population is vividly replaced by the frame index. We calculate the fitness based on the current frame, breed a new generation selectively, and use the new population to optimize continually based on the next frame. After several frames, the parameters can be optimized, and an ideal reconstructed frame can then be steadily output. Numerous experiments demonstrate the effectiveness and advantages of the proposed method.

A Recursive Decomposition Method for Large Scale Continuous Optimization

Cooperative co-evolution (CC) is an evolutionary computation framework that can be used to solve high-dimensional optimization problems via a “divide-and-conquer” mechanism. However, the main challenge when using this framework lies in problem decomposition. That is, deciding how to allocate decision variables to a particular subproblem, especially interacting decision variables. Existing decomposition methods are typically computationally expensive. In this paper, we propose a new decomposition method, which we call recursive differential grouping (RDG), by considering the interaction between decision variables based on nonlinearity detection. RDG recursively examines the interaction between a selected decision variable and the remaining variables, placing all interacting decision variables into the same subproblem. We use analytical methods to show that RDG can be used to efficiently decompose a problem, without explicitly examining all pairwise variable interactions. We evaluated the efficacy of the RDG method using large scale benchmark optimization problems. Numerical simulation experiments showed that RDG greatly improved the efficiency of problem decomposition in terms of time complexity. Significantly, when RDG was embedded in a CC framework, the optimization results were better than results from seven other decomposition methods.

Heuristic and metaheuristic solutions of pickup and delivery problem for self-driving taxi routing

Self-driving car automated control problem includes automated routing problem. This paper addresses the self-driving taxi routing problem formalized as the Pickup and Delivery problem (PDP). We use small moves technique as the basis for our evolutionary computation framework that considers small moves as mutation operations. We introduce strategies, which determine vector of probabilities to apply small moves on each step. Also we introduce policies, which determine how strategies evolve during the algorithm runtime. Finally, we also apply meta-learning to select the best strategy and policy for a given dataset. We test suggested methods with several genetic schemes. The best performance is shown by strategies that apply small moves with probability depending on its success rate and/or complexity. It outperforms Hill Climbing as well as a constraint satisfaction problem solver adopted for PDP.