Hybrid Evolutionary Algorithm Research Articles

Controlling Sequential Hybrid Evolutionary Algorithm by Q-Learning [Research Frontier] [Research Frontier

Many state-of-the-art evolutionary algorithms (EAs) can be categorized as sequential hybrid EAs, in which various EAs are sequentially executed. The timing to switch from one EA to another is critical to the performance of the hybrid EA because the switching time determines the allocation of computational resources and thereby it helps balance exploration and exploitation. In this article, a framework for adaptive parameter control for hybrid EAs is proposed, in which the switching time is controlled by a learned agent rather than a manually designed scheme. First the framework is applied to an adaptive differential evolution algorithm, LSHADE, to control when to use the scheme to reduce the population. Then the framework is applied to the algorithm that won the CEC 2018 competition, i.e., the hybrid sampling evolution strategy (HSES), to control when to switch from the univariate sampling phase to the Covariance Matrix Adaptation Evolution Strategy phase. The agents for parameter control in LSHADE and HSES are trained by using Q-learning and deep Q-learning to obtain the learned algorithms Q-LSHADE and DQ-HSES. The results of experiments on the CEC 2014 and 2018 test suites show that the learned algorithms significantly outperform their counterparts and some state-of-the-art EAs.

Spatiotemporal distribution and prediction of chlorophyll-a in Ulansuhai lake from an arid area of China

Lake Ulansuhai, a typical shallow lake in an arid area that is economically and ecologically important along the Yellow River, is currently eutrophic. Long-term (2010–2020) data on chlorophyll-a, nutrient, and environmental factors were obtained from three Lake Ulansuhai monitoring stations. The temporal and spatial distribution characteristics of Chl-a were analyzed. Additionally, a hybrid evolutionary algorithm was established to simulate and predict Chl-a, and sensitivity analysis revealed the interaction between environmental factors and eutrophication. The results indicated that (1) the seasonal variation of eutrophication showed an obvious trend of spring > summer > autumn > winter, and the concentration of Chl-a in the inlet was significantly higher than that in the outlet; (2) The inlet, center, and outlet of Ulansuhai Lake are satisfactorily affected by HEA in the best suited method. The fitting coefficients (R2) of the optimal models were 0.58, 0.59, and 0.62 for the three monitoring stations, and the root mean square errors (RMSE) were 3.89, 3.21, and 3.56, respectively; (3) under certain range and threshold conditions, Chl-a increased with the increase of permanganate index, water temperature, dissolved oxygen concentration, and ammonia nitrogen concentration, but decreased with the increase of water depth, Secchi disk depth, pH, and fluoride concentration. The results indicate that the HEA can simulate and predict the dynamics of Chl-a, and identify and quantify the relationships between eutrophication and the threshold data. The research results provide theoretical basis and technical support for the prediction and have great significance for the improvement of water quality and environmental protection in arid and semi-arid inland lakes.

Stochastic Flexible Power System Expansion Planning, Based on the Demand Response Considering Consumption and Generation Uncertainties

This paper presents the generation and transmission expansion planning (GTEP) considering the switched capacitive banks (SCBs) allocation in the power system, including the demand response program (DRP). This scheme is based on the system flexibility. The objective function of the scheme minimizes the expected planning cost that is equaled to the summation of the total construction costs of the SCBs, the generation units (GUs) and the transmission lines (TLs), and the operating cost of the GUs. It is concerned with the AC power flow constraints, the planning-operation model of the mentioned elements, the DRP operation formulation, and the operating and flexibility limits of the network. In the following, the scenario-based stochastic programming is used to model the uncertainty parameters, such as the load and renewable power of wind farms. Then, the hybrid evolutionary algorithm, based on the combination of the crow search algorithm and the grey wolf optimizer, is used to determine the optimal point with the approximate unique solution. Finally, the scheme is applied on the transmission networks, the numerical results confirm the capabilities of the proposed scheme in simultaneously improving the flexibility, operation, and economic situation of the transmission network, so that the hybrid algorithm achieves the optimal solution in a shorter computation time, compared with the non-hybrid algorithms. This algorithm has a low standard deviation of about 92% in the final response. The proposed scheme with the optimal planning of the lines, sources, and capacitor banks, together with the optimal operation of the DRP succeeded in improving the energy loss and the voltage deviation by about 30–36% and 25–30%, compared with those of the power flow studies.

Kernel-based hybrid multi-objective optimization algorithm (KHMO)

A new hybrid evolutionary algorithm named Kernel-based Hybrid Multi-Objective Optimization Algorithm (KHMO) is introduced in this work. The main novelty of this algorithm is its updating rule based on the concept of a reproducing kernel in order to properly approximate the numerical gradient. In addition, a novel numerical strategy based on the computation of a normal vector to determine a searching target direction that guides the nondominanted solutions into more propitious regions is also introduced for the first time. IGD and HV metrics are used to evaluate the performance of KHMO algorithm against other competitive multi-objective methods such as MRBFO, FDEA-I, VMEF, BiasMOSaDE and GAMODE tested on ZDT and CEC 2009 problems. Moreover, the Wilcoxon test is carried out as a comparative statistical analysis in order to detect if a significant difference exists with respect to the rest of the algorithms. The corresponding numerical results showed that the introduced KHMO algorithm gains the best overall performance among all the compared algorithms in terms of diversity, coverage and convergence in most problems. Moreover, IGD and HV curves reveal a respectable convergence rate since nearly 200 iterations are enough to offer quality solutions.

Accessing complex reconstructed material structures with hybrid global optimization accelerated via on-the-fly machine learning.

The complex reconstructed structure of materials can be revealed by global optimization. This paper describes a hybrid evolutionary algorithm (HEA) that combines differential evolution and genetic algorithms with a multi-tribe framework. An on-the-fly machine learning calculator is adopted to expedite the identification of low-lying structures. With a superior performance to other well-established methods, we further demonstrate its efficacy by optimizing the complex oxidized surface of Pt/Pd/Cu with different facets under (4 × 4) periodicity. The obtained structures are consistent with experimental results and are energetically lower than the previously presented model.

A new hybrid evolutionary algorithm for multi-objective optimal power flow in an integrated WE, PV, and PEV power system

• A novel hybrid multi-objective evolutionary algorithm (MOEA) was developed based on the decomposition and invasive weed optimization (IWO) techniques for solving the MOOPF problem. • The conventional MOOPF problem was modified by taking stochastic WE, PV, and PEV powers into consideration. • Modeling the stochastic nature of WE, PV, and PEV sources was done using probability distribution functions (PDFs) and calculating its uncertainty cost was done with Monte Carlo simulations. • Using an effective constraint handling method (CHM) known as the superiority of feasible solution (SF) was done to address difficult constraints. • The one-way analysis of variance (ANOVA) test was deployed to give a certain level of confidence to our study. This study proposes a novel hybrid multi-objective evolutionary algorithm (MOEA) based on decomposition and invasive weed optimization (IWO) algorithm for the optimal power flow (OPF) problem in transmission networks. The conventional OPF is modified as a stochastic OPF with the integration of WE, PV, and PEV systems uncertainty. This paper proposes a new constraint-handling method (CHM) that adds the penalty adaptively and avoids parameter reliance on penalty computation. The selection features in the IWO method are deployed to improve the diversity of the proposed method. The OPF problem is represented as a multi-objective optimization (MOO) problem using four objectives: generation cost, emission, power loss, and voltage variation. The generation costs of WE, PV, and PEV sources are evaluated using Monte Carlo simulations to mitigate the whole cost and the influence of intermittency of these systems is investigated in terms of affordability and feasibility. Weibull, lognormal, and normal probability distribution functions (PDFs) are employed to define the uncertainty of WE, PV, and PEV sources respectively. The suggested method's viability is evaluated on IEEE 57 and IEEE 118- bus systems under all conceivable scenarios. In addition, one-way ANOVA test, a statistical approach, is used to evaluate the superiority of the suggested algorithm.

A Bi-Objective Lane Reservation Problem Considering Dynamic Traffic Flow

Lane reservation strategies are widely used to ensure the right of way for eco-friendly vehicles and encourage people to green their commute. In most lane reservation problems (LRPs), the parameters underlying the system traffic conditions (e.g., vehicle speed and traffic flow) cannot be effectively specified. This paper addresses a new dynamic lane reservation problem (DLRP), which aims to optimize lanes that need to be reserved in different time periods based on trajectory data for existing transit network optimization. For passengers in reserved lanes, their travel time is minimized to guarantee traffic priority. Considering that the reserved lanes cause travel time growth on regular lanes, an improved multiobjective mixed integer nonlinear programming (MINLP) is established to minimize the delay. The problem complexity of this paper is NP-hard. This paper applies a preprocessing method for the actual traffic flow data analysis for link travel time calculation. We developed a hybrid evolutionary algorithm decomposing a multiobjective optimization problem (MOP) to a collection of simple MOPs. These subproblems are collaboratively solved. The hybrid crossover strategy exploits the advantages of different crossover operators for a better performance. The experimental results of MOEA/D and NSGA-II with standard test functions show that the proposed algorithm can improve the convergence and distribution of the results. Through numerous analyses and calculations of instances, the proposed algorithm is proven to be effective.

Multi-stage hybrid evolutionary algorithm for multiobjective distributed fuzzy flow-shop scheduling problem.

In the current global cooperative production mode, the distributed fuzzy flow-shop scheduling problem (DFFSP) has attracted much attention because it takes the uncertain factors in the actual flow-shop scheduling problem into account. This paper investigates a multi-stage hybrid evolutionary algorithm with sequence difference-based differential evolution (MSHEA-SDDE) for the minimization of fuzzy completion time and fuzzy total flow time. MSHEA-SDDE balances the convergence and distribution performance of the algorithm at different stages. In the first stage, the hybrid sampling strategy makes the population rapidly converge toward the Pareto front (PF) in multiple directions. In the second stage, the sequence difference-based differential evolution (SDDE) is used to speed up the convergence speed to improve the convergence performance. In the last stage, the evolutional direction of SDDE is changed to guide individuals to search the local area of the PF, thereby further improving the convergence and distribution performance. The results of experiments show that the performance of MSHEA-SDDE is superior to the classical comparison algorithms in terms of solving the DFFSP.

Multi-Objective Optimal Siting and Sizing of Distributed Generators and Shunt Capacitors Considering the Effect of Voltage-Dependent Nonlinear Load Models

Load modeling is essential to distribution system analysis, planning, and control. Therefore, in this work, effect of non-linear load models has been considered for the optimal site and size of DG and SC allocation. A new and efficient modified branch and bus ordering-based forward-backward load flow method has been applied to solve load flow problem in radial distribution system. Recently implemented several state-of-the-art evolutionary algorithms (EAs) are employed to solve optimal site and size of DG and SC allocation problems and it is shown that performance of multi-operator/multimethod is better than other algorithms that are based on a single operator and/or algorithm. Therefore, a new hybrid EA based on various state-of-the-art operators such as GA, DE, and PSO is designed and applied to solve optimal site and size of DG and SC allocation problems. Various technical objective functions (index of active and reactive power loss and voltage deviation index) are considered to show the impacts of non-linear load models. From the simulation results, it is shown that DG and SC allocation problem is multi-objective. Therefore, further weighted sum multi-objective technical, economic, and environmental functions are formulated to find the solution to DG and SC allocation problems. The gathered results demonstrate that the proposed methodology significantly minimizes the cost of energy supplied by grid, total operating cost, and active and reactive power losses. Consequently, it can be stated that the suggested methodology has considerable economic and technological benefits and may be used to address many optimization issues in various distribution networks.

Surrogate-assisted hybrid evolutionary algorithm with local estimation of distribution for expensive mixed-variable optimization problems

Some real-world design optimization problems can be formulated as expensive mixed-variable optimization problems (EMVOPs), which involve both continuous and discrete decision variables and expensive function evaluations. The main challenges for solving EMVOPs are the handling of mixed variables, limited number of function evaluations and multiple disconnected regions in the search space. In this work, we propose a novel algorithm with global and local search strategies for improving the search ability on disconnected regions, and it only consumes hundreds of function evaluations. The global module employs hybrid evolutionary operators and a Gower distance based surrogate model for handling mixed variables. The local estimation of distributions in different local regions performs in a competitive switching way to combine their advantages, and local surrogate models trained with selected samples improve the accuracy of approximated evaluations for the locally generated solutions. In the late stage, a local continuous search module is executed for refining the continuous decision variables. Verification results on the artificial benchmarks demonstrate that our proposed algorithm is competitive and works effectively. To verify the practicability of the algorithm, it is applied on a convolutional neural network hyperparameter optimization problems and obtains satisfactory results.

A novel model order reduction-based two-stage damage detection paradigm for trusses using time–history acceleration

This paper presents a novel model order reduction (MOR)-based two-stage damage detection method for trusses employing time-series acceleration measured by limited sensors. In the first step, an acceleration-based strain energy indicator (ASEI) is newly proposed to evaluate the most doubtfully damaged candidates. Accordingly, the number of design variables defined in an inverse optimization problem of the second phase is reduced significantly. The location and severity of damaged members are then determined by minimizing an objective function with a newly added dynamic penalty parameter to achieve a faster convergence speed and better optimal solutions. Since acceleration signals are incompletely measured at limited sensors, the second-order Neumann series expansion (SNSE) is firstly employed to condense the proportionally damped trusses for inferring the unmeasured time–history information. An adaptive hybrid evolutionary firefly algorithm (AHEFA) is utilized to resolve the above inverse optimization problem. Four numerical examples of 2D and 3D trusses with various damage scenarios including noises are examined to demonstrate the reliability of the proposed methodology. Attained outcomes indicate that the suggested paradigm can reliably diagnose both multidamage sites and extents of trusses with only relatively short time histories and a few measurement sensors.

Mathematical formulation and a hybrid evolution algorithm for solving an extended row facility layout problem of a dynamic manufacturing system

• An extended row facility layout problem is investigated considering dynamic manufacturing environment. • Specific row layout changes over multiple periods and a mixed integer linear programming model is built for it. • A hybrid evolution approach combining linear programming and evolutionary operations is developed. • A special decoding approach for the extended row facility layout problem is designed. • Comparison with other methods indicates that the proposed algorithm achieves highly preferable results. In this study, a dynamic extended row facility layout problem is investigated where the dynamic manufacturing process is considered. This problem involves the rearrangement cost when the location of the department changes over adjacent periods and it is aimed at minimising the sum of total material handling and rearrangement costs. The specific number of rows in each period is not predetermined in this study. Based on the product processing in practice, the essential clearance between departments and floor boundaries is also considered when calculating relative positions and exact locations of departments. A mixed-integer linear programming model is established for this problem. A hybrid evolution algorithm (HEA) that combines the exhaustive pairwise swap and variation operations is proposed to resolve this problem. First, a decoding strategy is designed to represent the department sequence and the optimal exact locations are solved using the linear programming method. Second, an exhaustive pairwise swap operation is addressed to search for the neighbourhood space without repeated calculations. This operation terminates when all swap pairs run out or preferable solution is found. Therefore, the variation operation starts immediately to switch to a new neighbourhood space. The HEA is tested with many instances and its good performance has been proven by comparing its results with existing solutions of fixed-row, flexible-row facility layout problems, and results of exact approach (CPLEX).

A Coevolutionary Algorithm for Cooperative Platoon Formation of Connected and Automated Vehicles

This paper proposes a coevolutionary algorithm to optimize longitudinal trajectories of multiple vehicles with an energy-aware non-linear objective during the cooperative platoon formation process. In this work, an adaptive encoding scheme is adopted to represent trajectories as knot vectors of parametric cubic splines, and therefore the original problem is reformulated into a constrained numerical optimization version. The number of knots can be adjusted to trade-off between the shape flexibility and computation efficiency. Further, the proposed coevolutionary algorithm decomposes the initially high-dimensional problem into smaller subproblems, significantly reducing the complexity. A hybrid evolutionary algorithm is developed as an optimizer for subproblems. Additionally, a branch-and-bound strategy and a Tabu search component are integrated into the steepest ascent hill-climbing algorithm to speed up convergences within the local exploitation phase. Numerical experiments are conducted on extensive scenarios with different platooning sizes and initial separations. Experimental results indicate the superiority of the proposed approach in optimality and stability with reasonable sub-second computation time for real-life applications.

A Hierarchical Approach to Intelligent Mission Planning for Heterogeneous Fleets of Autonomous Underwater Vehicles

The rapid development of marine robotic technology in recent decades has resulted in significant improvements in the self-sufficiency of autonomous underwater vehicles (AUVs). However, simple scenario-based approaches are no longer sufficient when it comes to ensuring the efficient interaction of multiple autonomous vehicles in complex dynamic missions. The necessity to respond cooperatively to constant changes under severe operating constraints, such as energy or communication limitations, results in the challenge of developing intelligent adaptive approaches for planning and organizing group activities. The current study presents a novel hierarchical approach to the group control system designed for large heterogeneous fleets of AUVs. The high-level core of the approach is rendezvous-based mission planning and is aimed to effectively decompose the mission, ensure regular communication, and schedule AUVs recharging activities. The high-level planning problem is formulated as an original acyclic variation of the inverse shift scheduling problem, which is NP-hard. Since regular schedule adjustments are supposed to be made by the robots themselves right in the course of the mission, a meta-heuristic hybrid evolutionary algorithm is developed to construct feasible sub-optimal solutions in a short time. The high efficiency of the proposed approach is shown through a series of computational experiments.

A multi‐objective hybrid evolutionary search algorithm for parallel production line balancing problem including disassembly and assembly tasks

AbstractThis paper proposes a multi‐objective hybrid evolutionary search algorithm to simultaneously optimize the number of workstations, the idle index and the quantity of the production equipment required for the parallel production line balancing problem including disassembly and assembly tasks. The proposed algorithm uses the crossover operation and the mutation operation to generate neighborhood individuals of the current evolutionary population. Based on the mixed population composed of the current evolutionary population and the newly generated neighborhood population, a novel population update method is developed. The individuals of the new generation population are composed of Pareto dominant individuals filtered from the mixed population and several individuals randomly selected from the dominated individuals. Then the local search strategy based on simulated annealing operation is applied to the new generation of population individuals to overcome the defect of premature convergence of the proposed algorithm. By solving test examples of different scales and comparing them with a variety of classical algorithms, the advantages in the aspects of the diversity, the convergence, and the distribution of the solution results of the proposed algorithm are proved. Finally, the established mathematical model and the developed algorithm are applied to the parallel production line designed for the bogie remanufacturing production line of a railway freight car maintenance company. The advantages of the parallel production line in improving the production efficiency and reducing the production cost are identified.

An Evolutionary Algorithm to Optimise a Distributed UAV Swarm Formation System

In this article, we present a distributed robot 3D formation system optimally parameterised by a hybrid evolutionary algorithm (EA) in order to improve its efficiency and robustness. To achieve that, we first describe the novel distributed formation algorithm3 (DFA3), the proposed EA, and the two crossover operators to be tested. The EA hyperparameterisation is performed by using the irace package and the evaluation of the three case studies featuring three, five, and ten unmanned aerial vehicles (UAVs) is performed through realistic simulations by using ARGoS and ten scenarios evaluated in parallel to improve the robustness of the configurations calculated. The optimisation results, reported with statistical significance, and the validation performed on 270 unseen scenarios show that the use of a metaheuristic is imperative for such a complex problem despite some overfitting observed under certain circumstances. All in all, the UAV swarm self-organised itself to achieve stable formations in 95% of the scenarios studied with a plus/minus ten percent tolerance.

Scalable feature subset selection for big data using parallel hybrid evolutionary algorithm based wrapper under apache spark environment.

Extant sequential wrapper-based feature subset selection (FSS) algorithms are not scalable and yield poor performance when applied to big datasets. Hence, to circumvent these challenges, we propose parallel and distributed hybrid evolutionary algorithms (EAs) based wrappers under Apache Spark. We propose two hybrid EAs based on the Binary Differential Evolution (BDE), and Binary Threshold Accepting (BTA), namely, (i) Parallel Binary Differential Evolution and Threshold Accepting (PB-DETA), where BDE and BTA work in tandem in every iteration, and (ii) its ablation variant, Parallel Binary Threshold Accepting and Differential Evolution (PB-TADE). Here, BTA is invoked to enhance the search capability and avoid premature convergence of BDE. For comparison purposes, we also parallelized two state-of-the-art algorithms: adaptive DE (ADE) and permutation based DE (DE-FSPM), and named them PB-ADE and P-DE-FSPM respectively. Throughout, logistic regression (LR) is employed to compute the fitness function, namely, area under the receiver operator characteristic curve (AUC). The effectiveness of the proposed algorithms is tested over the five big datasets of varying dimensions. It is noteworthy that the PB-TADE turned out to be statistically significant than the rest. All the algorithms have shown the repeatability property. The proposed parallel model attained a speedup of 2.2–2.9. We also reported feature subset with high AUC and least cardinality.

A hybrid evolutionary multitask algorithm for the multiobjective vehicle routing problem with time windows

The vehicle routing problem with time windows (VRPTW), as an important logistic problem, has been widely investigated in recent decades. However, it is still a great challenge for most existing approaches to solve multiobjective VRPTWs (MOVRPTWs) with many conflicting objectives. In this study, a hybrid evolutionary multitask algorithm, termed HEMT, is proposed to address MOVRPTWs under the framework of evolutionary multitasking, where multiple MOVRPTWs are optimized simultaneously by leveraging the similarity between them. In particular, HEMT is characterized by three aspects: (1) an exploration stage with knowledge transfer (ESKT) is designed to globally explore the search space by transferring knowledge across different MOVRPTWs; (2) an exploitation stage with knowledge reuse (ESKR) is employed to further promote the quality of solutions by conducting local searches and reusing problem-specific knowledge; and (3) a tradeoff mechanism (TOM) is suggested to adaptively switch the search between ESKT and ESKR. Furthermore, to verify the efficacy of the proposed HEMT, four suites of new multitasking instances are generated based on 45 real-world MOVRPTW benchmarks under different multitask environments. The experimental results show that HEMT not only effectively solves multiple MOVRPTWs simultaneously but also achieves better performances than the traditional single-task counterparts.

An environment-driven hybrid evolutionary algorithm for dynamic multi-objective optimization problems

In dynamic multi-objective optimization problems, the environmental parameters may change over time, which makes the Pareto fronts shifting. To address the issue, a common idea is to track the moving Pareto front once an environmental change occurs. However, it might be hard to obtain the Pareto optimal solutions if the environment changes rapidly. Moreover, it may be costly to implement a new solution. By contrast, robust Pareto optimization over time provides a novel framework to find the robust solutions whose performance is acceptable for more than one environment, which not only saves the computational costs for tracking solutions, but also minimizes the cost for switching solutions. However, neither of the above two approaches can balance between the quality of the obtained non-dominated solutions and the computation cost. To address this issue, environment-driven hybrid dynamic multi-objective evolutionary optimization method is proposed, aiming to fully use strengths of TMO and RPOOT under various characteristics of environmental changes. Two indexes, i.e., the frequency and intensity of environmental changes, are first defined. Then, a criterion is presented based on the characteristics of dynamic environments and the switching cost of solutions, to select an appropriate optimization method in a given environment. The experimental results on a set of dynamic benchmark functions indicate that the proposed hybrid dynamic multi-objective evolutionary optimization method can choose the most rational method that meets the requirements of decision makers, and balance the convergence and robustness of the obtained non-dominated solutions.

Ships’ response strategies to port disruptions caused by hurricanes

This paper studies ships' operational proactive strategies during the entire process of port disruptions due to environmental factors such as hurricanes. For the first time, the incompleteness and uncertainty of information, and the characteristics of port disruptions are considered. The problem is modeled as a nonhomogeneous Markov decision process (MDP), where different operational actions, such as sailing speed adjustment, handling rate adjustment, and port sequence adjustment are considered. The rolling horizon decision-making approach and a PSO-GA-based hybrid evolutionary algorithm are combined to solve the model. The results show that compared to the strategy without planned schedule adjustments, cost-saving up to 78.34% can be achieved by simultaneously adjusting the sailing speed, cargo handling rate, and port sequence. Counter-intuitively, we find that observing more complete hurricane information or having accurate forecasts does not always conducive to better decision-making. We also find the integration of multiple response strategies yields synergistic benefits. These findings will deepen our understanding of the impact of port disruptions and provide theoretical support for the decision-making of ship's response strategies.