Journal of Heuristics最新文献

The reliable fixed-charge facility location problem extends the fixed-cost facility location problem by incorporating facility unreliability. This paper addresses a novel capacitated version of the reliable fixed-charge facility location problem, where the failure probability of each facility is site-dependent, differing from existing literature models. Additionally, facilities are assumed to have limited capacity for demand allocation, with the option to increase capacity to a predetermined value in case of supply chain disruptions. To solve this problem, we develop a non-linear mixed-integer programming formulation and present its linear version. Given the NP-hard nature of the problem, we propose a novel Relax-and-Fix heuristic for its solution. We evaluate the efficiency of the proposed algorithm by solving a variety of experimental instances with different network sizes. Results demonstrate that the Relax-and-Fix heuristic improves upper bounds for problem instances and achieves this within a shorter computational time. Furthermore, sensitivity analysis is conducted on capacity and failure probabilities, and relevant results are presented.

The p-median problem has a central importance in the context of location planning problems. An extended version of this problem is the capacitated p-median problem (CPMP) which is used for diverse applications in urban planning and medical care units location. Given its relevance and its practicality, in this paper we present a large neighborhood search (LNS) and a study of hyper-heuristics for the CPMP. We propose and analyze various destruction operators within the framework of LNS to efficiently explore diverse neighborhoods. An exact solver is used in the repair phase. Additionally, these operators are also used for the hyper-heuristics, which are high-level problem-independent solution approaches, to propose new low-level heuristics. We provide a comparison of the LNS and of the best performing hyper-heuristics with state-of-art approaches for this problem. The proposed solution methods provide a lower average GAP value compared to the state of the art and find better solutions for several instances. We also give a detailed analysis of the performance of the low-level heuristics and their impact on the instances with different structure. These results confirm the robust performance of hyper-heuristics for the CPMP, which have the advantage to allow the employment on similar scenarios with minor adjustments.

The Burrows-Wheeler Transform (BWT) is a string transformation technique widely used in areas such as bioinformatics and file compression. Many applications combine a run-length encoding (RLE) with the BWT in a way which preserves the ability to query the compressed data efficiently. However, these methods may not take full advantage of the compressibility of the BWT as they do not modify the alphabet ordering for the sorting step embedded in computing the BWT. Indeed, any such alteration of the alphabet ordering can have a considerable impact on the output of the BWT, in particular on the number of runs. For an alphabet $\Sigma$ containing $\sigma$ characters, the space of all alphabet orderings is of size $\sigma !$ . While for small alphabets an exhaustive investigation is possible, finding the optimal ordering for larger alphabets is not feasible. Therefore, there is a need for a more informed search strategy than brute-force sampling the entire space, which motivates a new heuristic approach. In this paper, we explore the non-trivial cases for the problem of minimizing the size of a run-length encoded BWT (RLBWT) via selecting a new ordering for the alphabet. We show that random sampling of the space of alphabet orderings usually gives sub-optimal orderings for compression and that a local search strategy can provide a large improvement in relatively few steps. We also inspect a selection of initial alphabet orderings, including ASCII, letter appearance, and letter frequency. While this alphabet ordering problem is computationally hard we demonstrate gain in compressibility.

Reducing the cost of operating and maintaining wind farms is essential for the economic viability of this renewable energy source. This study applies hyper-heuristics to design a maintenance policy that prescribes the best maintenance action in every possible situation. Genetic programming is used to construct a priority function that determines what maintenance activities to conduct and the sequence of maintenance activities if there are not enough resources to do all of them simultaneously. The priority function may take into account the health condition of the target turbine and its components, the characteristics of the corresponding maintenance work, the workload of the maintenance crew, the working condition of the whole wind farm and the possibilities provided by opportunistic maintenance. Empirical results using a simulation model of the wind farm demonstrate that the proposed model can construct maintenance policies that perform well both in training and test scenarios, which shows the practicability of the approach.

This work address a variant of the knapsack problem, known as the knapsack problem with forfeits, which has numerous applications. In this variant, a set of items and a conflict graph are given, and the objective is to identify a collection of items that adhere to the knapsack’s capacity while maximizing the total value of the items minus the penalties for conflicting items. We propose a novel heuristic for this problem based on the concepts of iterated local search, variable neighborhood descent, and tabu search. Our heuristic takes into account four neighborhood structures, and we introduce efficient data structures to explore them. Experimental results demonstrate that our approach outperforms the state-of-the-art algorithms in the literature. In particular, it delivers superior solutions within significantly shorter computation times across all benchmark instances. Additionally, this study includes an analysis of how the proposed data structures have influenced both the quality of the solutions and the execution time of the method.

Over the past decade, the usefulness of quantum annealing hardware for combinatorial optimization has been the subject of much debate. Thus far, experimental benchmarking studies have indicated that quantum annealing hardware does not provide an irrefutable performance gain over state-of-the-art optimization methods. However, as this hardware continues to evolve, each new iteration brings improved performance and warrants further benchmarking. To that end, this work conducts an optimization performance assessment of D-Wave Systems’ Advantage Performance Update computer, which can natively solve sparse unconstrained quadratic optimization problems with over 5,000 binary decision variables and 40,000 quadratic terms. We demonstrate that classes of contrived problems exist where this quantum annealer can provide run time benefits over a collection of established classical solution methods that represent the current state-of-the-art for benchmarking quantum annealing hardware. Although this work does not present strong evidence of an irrefutable performance benefit for this emerging optimization technology, it does exhibit encouraging progress, signaling the potential impacts on practical optimization tasks in the future.

Throughout the response phase of the disaster, the speedy restoration of transportation by reconnecting the nodes where the connection is broken is absolutely critical for evacuating civilians, providing clear access to hospitals, and distributing aid. Following a disaster, some roads in a disaster area might be closed to transportation. In reality, some roads can be blocked due to debris, and some of roads can be blocked by collapsing. In this model, different types of road unblocking methods are included, and each road can only be opened to access by a vehicle suitable for that method. So, different types of vehicles may be needed to repair the roads depending on the type of damage. In addition, fast-built bridges built both on land and over water are also used if necessary following a disaster. In problems of this nature, it is essential to restore the roads to enable the complete connectivity of the network such that all nodes can be reached by one another. In addition, it is also critical for the speedy reach of critical nodes, such as hospitals, and emergency disaster centers. This study aims to reduce the maximum time for connection and minimize the total time in which to reach critical nodes. For this purpose, we developed a bi-objective mathematical model that considers the multiple vehicle types that can repair different types of damages. Since the problem is NP-hard, two heuristic methods were developed, and the numerical results were presented. It has been observed that the local search algorithm gives better results than the hybrid algorithm. Additionally, different scenario data was produced. Numbers of unconnected components from 3 to 10 are solved with heuristic algorithms for test data containing 80 and 250 nodes, and real-life data containing 223 nodes and 391 edges are solved with heuristic algorithms for the number of unconnected components 6, 9, 12, and 15.

The adjacency graph is a structure used to model genomes in several rearrangement distance problems. In particular, most studies use properties of a maximum cycle packing of this graph to develop bounds and algorithms for rearrangement distance problems, such as the reversal distance, the reversal and transposition distance, and the double cut and join distance. When each genome has no repeated genes, there exists only one cycle packing for the graph. However, when each genome may have repeated genes, the problem of finding a maximum cycle packing for the adjacency graph (adjacency graph packing) is NP-hard. In this work, we develop a randomized greedy heuristic and a genetic algorithm heuristic for the adjacency graph packing problem for genomes with repeated genes and unequal gene content. We also propose new algorithms with simple implementation and good practical performance for reversal distance and reversal and transposition distance in genomes without repeated genes, which we combine with the heuristics to find solutions for the problems with repeated genes. We present experimental results and compare the application of these heuristics with the application of the MSOAR framework in rearrangement distance problems. Lastly, we apply our genetic algorithm heuristic to real genomic data to validate its practical use.

Discrete diversity optimization basically consists of selecting a subset of elements of a given set in such a way that the sum of their pairwise distances is maximized. Equity, on the other hand, refers to minimizing the difference between the maximum and the minimum distances in the subset of selected elements to balance their diversity. Both problems have been studied in the combinatorial optimization literature, but recently major drawbacks in their classic mathematical formulations have been identified. We propose new mathematical models to overcome these limitations, including multi-objective optimization, and heuristics to solve large-size instances of them. Specifically, we propose a matheuristic based on the CMSA framework for diversity and a GRASP heuristic for equity. Our extensive experimentation compares the original models with the new proposals by analyzing the solutions of our heuristics and those of the previous approaches, both from a single objective and a bi-objective paradigm. We also evaluate their quality with respect to the optimal solutions obtained with CPLEX, size permitting. Statistical analysis allows us to draw significant conclusions.

Methaheuristics (MHs) are techniques widely used for solving complex optimization problems. In recent years, the interest in combining MH and machine learning (ML) has grown. This integration can occur mainly in two ways: ML-in-MH and MH-in-ML. In the present work, we combine the techniques in both ways—ML-in-MH-in-ML, providing an approach in which ML is considered to improve the performance of an evolutionary algorithm (EA), whose solutions encode parameters of an ML model—artificial neural network (ANN). Our approach called TS(_{in})EA(_{in})ANN employs a reinforcement learning neighborhood (RLN) mutation based on Thompson sampling (TS). TS is a parameterless reinforcement learning method, used here to boost the EA performance. In the experiments, every candidate ANN solves a regression problem known as protein structure prediction deviation. We consider two protein datasets, one with 16,382 and the other with 45,730 samples. The results show that TS(_{in})EA(_{in})ANN performs significantly better than a canonical genetic algorithm (GA(_{in})ANN) and the evolutionary algorithm without reinforcement learning (EA(_{in})ANN). Analyses of the parameter’s frequency are also performed comparing the approaches. Finally, comparisons with the literature show that except for one particular case in the largest dataset, TS(_{in})EA(_{in})ANN outperforms other approaches considered the state of the art for the addressed datasets.