Hybrid Local Search Research Articles

A hybrid local search algorithm for the continuous energy-constrained scheduling problem

AbstractWe consider the continuous energy-constrained scheduling problem (CECSP). A set of jobs has to be processed on a continuous, shared resource. A schedule for a job consists of a start time, completion time, and a resource consumption profile. The goal is to find a schedule such that each job does not start before its release time, is completed before its deadline, satisfies its full resource requirement, and respects its lower and upper bounds on resource consumption during processing. The objective is to minimize the total weighted completion time. We present a hybrid local search approach, using simulated annealing and linear programming, and compare it to a mixed-integer linear programming (MILP) formulation. We show that the hybrid local search approach matches the MILP formulation in solution quality for small instances and is able to find a feasible solution for larger instances in reasonable time.

Knowledge-Guided Parallel Hybrid Local Search Algorithm for Solving Time-Dependent Agile Satellite Scheduling Problems

As satellite capabilities have evolved and new observation requirements have emerged, satellites have become essential tools in disaster relief, emergency monitoring, and other fields. However, the efficiency of satellite scheduling still needs to be enhanced. Learning and optimization are symmetrical processes of solving problems. Learning problem knowledge could provide efficient optimization strategies for solving problems. A knowledge-guided parallel hybrid local search algorithm (KG-PHLS) is proposed in this paper to solve time-dependent agile Earth observation satellite (AEOS) scheduling problems more efficiently. Firstly, the algorithm uses heuristic algorithms to generate initial solutions. Secondly, a knowledge-based parallel hybrid local search algorithm is employed to solve the problem in parallel. Meanwhile, data mining techniques are used to extract knowledge to guide the construction of new solutions. Finally, the proposed algorithm has demonstrated superior efficiency and computation time through simulations across multiple scenarios. Notably, compared to benchmark algorithms, the algorithm improves overall efficiency by approximately 7.4% and 8.9% in large-scale data scenarios while requiring only about 60.66% and 31.89% of the computation time of classic algorithms. Moreover, the proposed algorithm exhibits scalability to larger problem sizes.

WordBlitz: An Efficient Hard-Label Textual Adversarial Attack Method Jointly Leveraging Adversarial Transferability and Word Importance

Existing textual attacks mostly perturb keywords in sentences to generate adversarial examples by relying on the prediction confidence of victim models. In practice, attackers can only access the prediction label, meaning that the victim model can easily defend against such hard-label attacks by denying access based on the attack’s frequency. In this paper, we propose an efficient hard-label attack approach, called WordBlitz. First, based on the adversarial transferability, we train a substitute model to initialize the attack parameter set, including a candidate pool and two weight tables of keywords and candidate words. Then, adversarial examples are generated and optimized under the guidance of the two weight tables. During optimization, we design a hybrid local search algorithm with word importance to find the globally optimal solution while updating the two weight tables according to the attack results. Finally, the non-adversarial text generated during perturbation optimization is added to the training of the substitute model as data augmentation to improve the adversarial transferability. Experimental results show that WordBlitz surpasses the baseline in terms of better effectiveness, higher efficiency, and lower cost. Its efficiency is especially pronounced in scenarios with broader search spaces, and its attack success rate on a Chinese dataset is higher than on baselines.

Traveling of multiple salesmen to dynamically changing locations for satisfying multiple goals

Polymer grade scheduling, maritime surveillance, e-food delivery, e-commerce, and military tactics necessitate multiple agents (e.g., extruders, speed boats, salesmen) capable of visiting (or completing) dynamically changing locations (or tasks) in minimum time and distance. This study proposes a novel methodology based on clustering and local heuristic-based evolutionary algorithms to address the dynamic traveling salesman problem (TSP) and the dynamic multi-salesman problem with multiple objectives. The proposed algorithm is evaluated on 11 benchmark TSP problems and large-scale problems with up to 10,000 instances. The results show the superior performance of the proposed methodology called the dynamic two-stage evolutionary algorithm as compared to the dynamic hybrid local search evolutionary algorithm. Furthermore, the algorithm's applicability is illustrated through various scenarios involving up to four salesmen and three objectives with dynamically changing locations. To demonstrate real-world relevance, a maritime surveillance problem employing a helideck monitoring system is solved, wherein the objective is to minimize the patrolling route while visiting faulty vessels that threaten marine vessels. This study provides a general framework of TSP which finds application in several sectors, including planning and scheduling in chemical and manufacturing industries, the defense sector, and the e-commerce sector. Finally, the results showcase the effectiveness of the proposed methodology in solving the dynamic multiobjective, and multiple salesmen problem, which represents a more generalized version of the TSP.

Effective social spider optimization algorithms for distributed assembly permutation flowshop scheduling problem in automobile manufacturing supply chain

This paper presents a novel distributed assembly permutation flowshop scheduling problem (DAPFSP) based on practical problems in automobile production. Different from the existing research on DAPFSP, this study considers that each component of the final product is composed of more than one part. Components are processed in a set of identical components manufacturing factories and are assembled into products in the assembly factory. The integration of manufacturing processes is an important objective of Industry 4.0. For solving this problem with the minimum makespan criterion, we introduce a three-level representation and a novel initialization method. To enhance the search ability of the proposed algorithms, we design three local search methods and two restart procedures according to characteristics of the problem. Then, by incorporating the problem specific knowledge with the social spider optimization algorithm (SSO), we propose three SSO variants: the SSO with hybrid local search strategies (HSSO), the HSSO with restart procedures (HSSOR), and the HSSOR with self-adaptive selection probability (HSSORP). Finally, 810 extended instances based on the famous instances are used to test the proposed algorithms. In most cases, HSSOR performs the best, with an average comparison metric value of 0.158% across three termination conditions, while the average comparison metric value for the best comparison method is 2.446%, which is 15.481 times that of HSSOR. Numerical results demonstrate that the proposed algorithms can solve the problem efficiently.

An Hybrid Local Search for the Direct Aperture Optimisation Problem

Radiotherapy is a cancer treatment that uses high levels of radiation to destroy cancerous cells and shrink tumours while minimising harm to surrounding organs at risk (OARs). One of the techniques used in radiotherapy is Intensity Modulated Radiation Therapy (IMRT). Usually, the IMRT problem is approached sequentially, that is, we first need to determine the set of beam angles from which radiation will be delivered. Then, the radiation intensities for each selected beam angle are computed. Finally, the sequence of aperture shapes needed to deliver the computed treatment plan is generated. Unfortunately, the treatment plans generated by this approach have many apertures, which leads to longer treatment times. In contrast, the Direct Aperture Optimisation (DAO) problem considers constraints associated with the number of deliverable aperture shapes and physical constraints of the machine during the optimisation process of the intensities. The DAO approach generates, in general, better treatments with fewer apertures for IMRT. This is important because fewer apertures usually means shorter delivery times. In this work, we propose an hybrid local search strategy with mathematical programming to efficiently solve the DAO problem. We apply our proposed local search algorithm to a set of prostate cases, obtaining very competitive results.

Mixed-integer programming model and hybrid local search genetic algorithm for human–robot collaborative disassembly line balancing problem

ABSTRACT Human–robot collaborative technology maximises the advantages of the capabilities of humans and robots, and provides diverse operating scenarios for the remanufacturing industry. Accordingly, this paper proposes an innovative human–robot collaborative disassembly line balancing problem (HRC-DLBP). First, a mixed-integer programming (MIP) model is devised for the HRC-DLBP to minimise the number of workstations, smoothness index, and various costs. Second, a hybrid local search genetic algorithm (HLSGA) is developed to solve the proposed HRC-DLBP efficiently. According to the problem characteristics, a four-layer encoding and decoding strategy was constructed. The search mechanism of the local search operator was improved, and its search strategy was adjusted to suit the genetic algorithm structure better. Furthermore, the accuracy of the proposed MIP model and HLSGA is verified through two HRC-DLBP examples. Subsequently, three HRC-DLBP examples are used to prove that the HLSGA is superior to five other excellent algorithms. The case of the two-sided disassembly line problem reported in the literature is also solved using the HLSGA. The results are found to be significantly better than the reported outputs of the improved whale optimisation algorithm. Besides, HLSGA also outperforms the results reported in the literature in solving EOL state-oriented DLBP. Finally, the HLSGA is applied to a power battery disassembly problem, and several optimal allocation schemes are obtained.

Integrating machine layout, transporter allocation and worker assignment into job-shop scheduling solved by an improved non-dominated sorting genetic algorithm

To meet the everchanging production demand and close to the actual scheduling environment, a multi-objective job-shop scheduling problem (MOJSP) with multiple resource constraints is investigated. Specifically, an integrated mathematical model is constructed based on a job-shop scheduling problem (JSP) considering machine layout rearrangement, transporter allocation with capacity limitation, and worker assignment with skill variance to simultaneously minimize the exit time, labor cost, worker workload difference, and transportation time. To tackle the concerned problem, an improved non-dominated sorting genetic algorithm with a hybrid local search (INSGA-HLS) is designed. In the numerical simulation, a test dataset is first constructed according to the literature. Second, an orthogonal experiment is utilized to find the best combination of key parameters for INSGA-HLS. Third, the exploitation competence of the proposed hybrid local search (HLS) is verified. Then, the superiority of the designed INSGA-HLS algorithm is demonstrated by comparing it with the other intelligent algorithms. Thereafter, the influence of three considered factors on the integrated scheduling problem is illustrated: (1) flexible machine layout rearrangement may reduce the handling time, and improve production efficiency to some extent; (2) transporter capacity limitation can prolong the transport time, which extends the exit time; and (3) worker assignment with skill variance can not only improve productivity and reduce the labor cost, but also narrow the workload difference to improve employee satisfaction. Finally, an actual production line application further verifies the practicability and benefit of the model, and the managerial significance is analyzed.

Hybrid Local Search Algorithm for Optimization Route of Travelling Salesman Problem

Hybrid Local Search Algorithm for Optimization Route of Travelling Salesman Problem

Reinforcement learning based coarse-to-fine search for the maximum [formula omitted]-plex problem

The maximum k-plex problem is a computationally challenging problem, which emerged from graph-theoretic social network studies in the context of community detection. As a relaxation of the traditional maximum clique problem, intended to identify cohesive structures within a graph that are not necessarily fully connected, the aim of this problem is to find the largest set of vertices in a graph in which all vertices are not adjacent to at most k vertices within this set. This paper introduces an effective hybrid local search algorithm for solving the maximum k-plex problem that combines a stochastic local search method and a reinforcement learning strategy. The proposed approach includes a number of distinguishing features, including: a new coarse-to-fine strategy to balance the search process, a distance-and-quality reward for actions, and a parameter control mechanism based on reinforcement learning. We conduct a computational analysis of the key components of the proposed algorithm, assessing their impact on the overall performance. Extensive experiments for the maximum k-plex problem (k=2,3,4,5) on 80 benchmark instances from the second DIMACS Challenge demonstrate that the proposed approach competes favourably with the state-of-the-art algorithms from the literature.

The objective that freed me: a multi-objective local search approach for continuous single-objective optimization

Single-objective continuous optimization can be challenging, especially when dealing with multimodal problems. This work sheds light on the effects that multi-objective optimization may have in the single-objective space. For this purpose, we examine the inner mechanisms of the recently developed sophisticated local search procedure SOMOGSA. This method solves multimodal single-objective continuous optimization problems based on first expanding the problem with an additional objective (e.g., a sphere function) to the bi-objective domain and subsequently exploiting local structures of the resulting landscapes. Our study particularly focuses on the sensitivity of this multiobjectivization approach w.r.t. (1) the parametrization of the artificial second objective, as well as (2) the position of the initial starting points in the search space. As SOMOGSA is a modular framework for encapsulating local search, we integrate Nelder–Mead local search as optimizer in the respective module and compare the performance of the resulting hybrid local search to its original single-objective counterpart. We show that the SOMOGSA framework can significantly boost local search by multiobjectivization. Hence, combined with more sophisticated local search and metaheuristics, this may help solve highly multimodal optimization problems in the future.

An Effective Iterated Greedy Algorithm for a Robust Distributed Permutation Flowshop Problem With Carryover Sequence-Dependent Setup Time

A new scheduling problem, the distributed permutation flowshop scheduling problem with uncertain processing times and carryover sequence-dependent setup time (DPUC), is addressed. The DPUC is an important application problem in modern electronics manufacturing. A robust model is established for the DPUC with makespan criterion. A counter-intuitive paradox is found, that is, adding a new job to one of the production lines can reduce the completion time of the production line. Two acceleration methods are provided to save computational efforts. An iterated greedy algorithm called IG_FS is proposed to solve the DPUC. A heuristic based on the well-known NEH is proposed to generate the initial solution for the IG_FS. In the destruction phase of the IG_FS, dynamic sizes based on both adaptability and randomness are provided to improve the exploration capability. During the local search phase of the IG_FS, a hybrid local search method consisting of shift and swap operators is presented to exploit more diverse search areas. Extensive experiments show that the proposed IG_FS performs significantly better than the six competing algorithms adapted from the closely related scheduling literature.

A hybrid multi-objective evolutionary algorithm for solving an adaptive flexible job-shop rescheduling problem with real-time order acceptance and condition-based preventive maintenance

Production scheduling and maintenance planning are two of the most important tasks in the modern manufacturing workshop. Meanwhile, due to the dynamic order arrival and real-time machine monitoring information updating, the integrated optimization of them becoming more complex and meaningful. Therefore, this study intends to address an adaptive flexible job-shop rescheduling problem with real-time order acceptance (ROA) and condition-based preventive maintenance (CBPM). More precisely, the main innovative works are described as follows: (1) a CBPM policy with both imperfect preventive maintenance (PM) and four inspection strategies is designed to find the optimal maintenance planning for each production machine; (2) a multi-objective optimization model is developed for the concerned problem; and (3) a hybrid multi-objective evolutionary algorithm (HMOEA) with hybrid initialization method, hybrid local search operators and adaptive rescheduling strategies is proposed. In the numerical simulation, the performance and competitiveness of the proposed CBPM policy are first demonstrated by comparing with other maintenance policies. Second, the effectiveness and superiority of parameter setting, order sorting rules, improved operators and overall performance of the proposed algorithm are verified by internal analysis of the algorithm. Third, an adaptive rescheduling strategy pool is constructed by running three rescheduling strategies on all rescheduling scenarios. Finally, a comprehensive sensitivity analysis is performed to illustrate the impact of several critical parameters on the adaptive rescheduling problem, and the results and comparisons show that the proposed HMOEA algorithm and order acceptance strategy have good robustness in most parameters.

Stochastic multi-objective optimization approaches in a real-world oil field waterflood management

Multi-objective differential evolution with hybrid local dynamic search algorithm (HMODE-DLS) and non-dominated sorting genetic algorithm (NSGA-II) are applied successfully in a reservoir sector model of a Middle Eastern real-world oil field under waterflood development. The Pareto solutions of the two objectives of maximum oil production and minimum water production obtained by the two algorithms are analyzed and compared in this study. The sector model consists of four oil producers and three water injectors and it is run for ten years with twenty time-steps of six months each. A total of 140 decision variables are defined as bottom-hole pressures for the four producers and water injection rates for the three water injectors at each time step (i.e. 7 variables by 20 time-steps). The optimal Pareto solution is obtained at 12,600 function evaluations by HMODE-DLS with 70% convergence improvement compared to 42,000 objective evaluations by NSGA-II. Results show that NSGA-II generated more optimum solutions in the lower range (i.e. less than 480 thousand m3 (Mm3) total oil production) whereas, the hybrid MODE-DLS has a more optimum solution in the higher range of the Pareto front. The net flow method (NFM) is usedin this study to select the best optimal solutions from the Pareto front solutions based on the weight of the objective function chosen by the decision-maker. By using the entropy weight criteria, the best optimal solution of high oil production determined is 540.45 Mm3 at the HMODE-DLS Pareto front. On the other hand, the best optimal solution of low oil production obtained is 469.93 Mm3 at the NSGA-II Pareto front. The combined optimal Pareto fronts obtained by NSGA-II and HMODE-DLS algorithms offer good spread and diverse solutions for the decision-maker to select and operate the field based on the operation conditions and constraints.

A hybrid local search algorithm for minimum dominating set problems

Minimum dominating set (MDS) is a well-known NP-hard fundamental graph theory problem having many applications such as mining social networks and bioinformatics. MDS seeks for the minimum subset of vertices in which every vertex not in the selected subset is adjacent to at least one vertex of this subset. In this study, we consider MDS and a complex variant of MDS known as the minimum positive influence dominating set (MPIDS). In MPIDS, the aim is to identify a subset of vertices where each vertex of a graph must be dominated by at least half of its neighbors. To solve these problems, we propose a two-stage hybrid local search algorithm. In the first stage, we propose an adaptive information content-based local search algorithm as an exploratory procedure. This algorithm focuses on generating promising solutions in different areas of the solution space using the problem search history. Moreover, we introduce information content strategy-based neighborhood structure and tolerance-based features to evolve high-quality and diverse solutions. For the second stage, we propose a gain-based deterministic local search algorithm as an exploitation procedure. It navigates feasible neighborhood areas to improve the solution quality. We conducted extensive analysis to evaluate the impact of the proposed components. The proposed algorithm produced very good results and generalized well overall instances of both MDS and MPIDS. Compared to the literature, the proposed algorithm outperformed other algorithms in most of the tested instances.

A hash map-based memetic algorithm for the distributed permutation flowshop scheduling problem with preventive maintenance to minimize total flowtime

Distributed scheduling problems have been an active research topic due to their close connection with the multifactory production environment in today’s decentralized economy. In this paper, the distributed permutation flowshop scheduling problem with preventive maintenance is studied with the goal of minimizing the total flowtime. A mathematical model and a hash map-based algorithm are presented to tackle the problem. The Nawaz–Enscore–Ham heuristic is improved and incorporated with the distributed Liu–Reeves heuristic to provide a promising initial solution in a finite number of steps. The proposed algorithm employs the data structure of the hash map to store all the candidate solutions in the form of key–value pairs. The selection, crossover, and mutation operators are also modified to expand the scope of exploration in the discrete domain. The local search hybridizes the job insertion operator and the job swap operator to further improve the obtained offspring solutions and utilizes the characteristic of the population hash map to reduce the efforts of solution evaluations. The candidate solution obtained by the local search is hashed by a rotating hash method and then placed at the node with the corresponding hash code. A series of experiments were conducted to verify the effectiveness of the hybrid local search operators and hash map strategy. Computational results indicate that the multiple strategies not only help to escape from the local optimization but also improve the computational efficiency. When compared with 10 state-of-the-art algorithms, the proposed algorithm generates an average relative percentage deviation of 0.134%, which is significant improvement.

A Local Search Q-learning Based Approach for Multi-period Vehicle Redeployment

The problem of vehicle redeployment in multi-period systems is studied in this paper. It is a sequential decision making problem that involves optimizing an objective over some planning horizon composed of multiple periods. In this work, the problem is handled with a hybrid local search and Q-learning algorithm. It integrates local search into the Dyna-Q algorithm and combines the system’s tactical and operational decision levels to find a multi-period policy for sharing vehicles between stations while optimizing the objective. Tests were carried out on synthetic and real instances, and the obtained results show how local search integration can improve the basic Dyna-Q algorithm.

Grid-Based Hybrid Genetic Approach to Relaxed Flexible Flow Shop with Sequence-Dependent Setup Times

In this paper, a hybrid genetic algorithm implemented in a grid environment to solve hard instances of the flexible flow shop scheduling problem with sequence-dependent setup times is introduced. The genetic algorithm takes advantage of the distributed computing power on the grid to apply a hybrid local search to each individual in the population and reach a near optimal solution in a reduced number of generations. Ant colony systems and simulated annealing are used to apply a combination of iterative and cooperative local searches, respectively. This algorithm is implemented using a master–slave scheme, where the master process distributes the population on the slave process and coordinates the communication on the computational grid elements. The experimental results point out that the proposed scheme obtains the upper bound in a broad set of test instances. Also, an efficiency analysis of the proposed algorithm indicates its competitive use of the computational resources of the grid.

An Ultra-Fast Power Prediction Method Based on Simplified LSSVM Hyperparameters Optimization for PV Power Smoothing

With existing power prediction algorithms, it is difficult to satisfy the requirements for prediction accuracy and time when PV output power fluctuates sharply within seconds, so this paper proposes a high-precision and ultra-fast PV power prediction algorithm. Firstly, in order to shorten the optimization time and improve the optimization accuracy, the single-iteration Gray Wolf Optimization (SiGWO) method is used to simplify the iteration process of the hyperparameters of Least Squares Support Vector Machine (LSSVM), and then the hybrid local search algorithm composed of Iterative Local Search (ILS) and Self-adaptive Differential Evolution (SaDE) is used to improve the accuracy of hyperparameters, so as to achieve high-precision and ultra-fast PV power prediction. The power prediction model is established, and the proposed algorithm is applied in a test experiment which can complete the power prediction within 3 s, and the RMSE is only 0.44%. Finally, combined with the PV-storage advanced smoothing control strategy, it is verified that the performance of the proposed algorithm can satisfy the system’s requirements for prediction accuracy and time under the condition of power mutation in a PV power generation system.

A population-based iterated greedy algorithm to minimize total flowtime for the distributed blocking flowshop scheduling problem

We consider the distributed blocking flowshop scheduling problem (DBFSP) which is a meaningful generalization of the blocking flowshop scheduling problem in the distributed production environment. The objective of minimizing the total flowtime is relevant and important in the current dynamic manufacturing environment, but, as far as we know, it has not been investigated in the DBFSP previously. In this paper, a population-based iterated greedy (PBIG) algorithm is proposed to solve the DBFSP with the total flowtime criterion, which takes the advantage of both the population-based search approach and the iterated greedy algorithm. First, an effective constructive heuristic is proposed by integrating two existing constructive approaches to initialize the population with a high level of quality and diversity. Second, three different procedures to generate the offspring solutions are tested for the effective exploration capability, each of which rationally combines the destruction, reconstruction and selection operator. Third, the insertion neighborhood and swap neighborhood are investigated to enhance the local exploitation ability and a hybrid local search procedure that utilizes simultaneously both the two neighborhoods are proposed. The comprehensive experimental evaluation based on a total of 720 well-known instances shows that the proposed algorithms outperform the existing effective algorithms at a significant margin.