A survey of adaptive large neighborhood search algorithms and applications

Mara, Setyo Tri Windras; Norcahyo, Rachmadi; Jodiawan, Panca; Lusiantoro, Luluk; Rifai, Achmad Pratama

doi:10.1016/j.cor.2022.105903

Cited by 29 publications

(4 citation statements)

References 269 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ALNS is a metaheuristic algorithm based on the idea of large neighborhood search [40,41]. It has received considerable attention in recent years and has been successfully applied to many problem areas [42], such as vehicle routing problems [43,44], scheduling problems [45], and pod storage assignment in RMFS warehouses [13]. The ALNS algorithm needs to design multiple destroy methods and repair methods, so applying ALNS to a new problem domain is difficult [13].…”

Section: Literature Reviewmentioning

confidence: 99%

Item Storage Assignment Problem in Robotic Mobile Fulfillment Systems With Nonempty Pods

Li,

Deng,

2024

IEEE Access

View full text Add to dashboard Cite

This study examines the item storage assignment problem in robotic mobile fulfillment systems (RMFSs), that determines the assignment of items to pods. Unlike previous studies on this problem that generally considered an empty warehouse, the current study addresses the problem in RMFSs with nonempty pods. The problem is formulated as a mixed-integer program with the goal of maximizing the correlation degree among items. On the basis of the characteristics of the problem, an adaptive large neighborhood search (ALNS) heuristic is designed for the solution. Numerical results show that the gap between the ALNS heuristic and the Gurobi solver is less than 0.32% in small-scale instances. In medium-and large-scale instances, the ALNS heuristic improves the objective value by at least 10% compared with the methods in literature, and it outperforms commonly used storage assignment policies for traditional warehouses (e.g., random, dedicated, and class-based storage) by more than 30%. The ALNS heuristic can also be applied to the special case where the initial state of the warehouse is empty. The computational study of the ALNS heuristic in all instances shows that it can effectively solve this problem by providing high-quality solutions and has good scalability.INDEX TERMS Robotic mobile fulfillment system, storage assignment, adaptive large neighborhood search, warehouse.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Item Storage Assignment Problem in Robotic Mobile Fulfillment Systems With Nonempty Pods

Li,

Deng,

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…2. Roulette wheel selection is commonly used to execute and adapt the destroy heuristic selection to determine the neighborhood (Mara et al 2022;Li et al 2021). During optimization, roulette wheel selection could greedily converge to poor choices due to the lack of exploration.…”

Section: Introductionmentioning

confidence: 99%

Adaptive Anytime Multi-Agent Path Finding Using Bandit-Based Large Neighborhood Search

Phan,

Huang,

Dilkina

et al. 2024

AAAI

View full text Add to dashboard Cite

Anytime multi-agent path finding (MAPF) is a promising approach to scalable path optimization in large-scale multi-agent systems. State-of-the-art anytime MAPF is based on Large Neighborhood Search (LNS), where a fast initial solution is iteratively optimized by destroying and repairing a fixed number of parts, i.e., the neighborhood of the solution, using randomized destroy heuristics and prioritized planning. Despite their recent success in various MAPF instances, current LNS-based approaches lack exploration and flexibility due to greedy optimization with a fixed neighborhood size which can lead to low-quality solutions in general. So far, these limitations have been addressed with extensive prior effort in tuning or offline machine learning beyond actual planning. In this paper, we focus on online learning in LNS and propose Bandit-based Adaptive LArge Neighborhood search Combined with Exploration (BALANCE). BALANCE uses a bi-level multi-armed bandit scheme to adapt the selection of destroy heuristics and neighborhood sizes on the fly during search. We evaluate BALANCE on multiple maps from the MAPF benchmark set and empirically demonstrate performance improvements of at least 50% compared to state-of-the-art anytime MAPF in large-scale scenarios. We find that Thompson Sampling performs particularly well compared to alternative multi-armed bandit algorithms.

show abstract

“…This allows for accepting less promising solutions in the search, aiming to break out of local optima and search potentially more promising areas of the search space. ALNS is a popular metaheuristic for solving largescale planning and scheduling problems, including routing and job scheduling problems (Mara et al 2022).…”

Section: Introductionmentioning

confidence: 99%

“…Additionally, the efficacy of operators is influenced by their parameter configurations. Existing configuration methods are mainly based on conventions, experience, ad hoc choices, and experimentation (Mara et al 2022), which may be suboptimal as they do not consider the problem-specific characteristics nor incorporate online information about the current search process dynamics. Additionally, experimentation faces limitations as the parameters interact, making grid or random search impractical for computationally expensive real-world problems.…”

Section: Introductionmentioning

confidence: 99%

Online Control of Adaptive Large Neighborhood Search Using Deep Reinforcement Learning

Reijnen,

Zhang,

Lau

et al. 2024

ICAPS

View full text Add to dashboard Cite

The Adaptive Large Neighborhood Search (ALNS) algorithm has shown considerable success in solving combinatorial optimization problems (COPs). Nonetheless, the performance of ALNS relies on the proper configuration of its selection and acceptance parameters, which is known to be a complex and resource-intensive task. To address this, we introduce a Deep Reinforcement Learning (DRL) based approach called DR-ALNS that selects operators, adjusts parameters, and controls the acceptance criterion throughout the search. The proposed method aims to learn, based on the state of the search, to configure ALNS for the next iteration to yield more effective solutions for the given optimization problem. We evaluate the proposed method on an orienteering problem with stochastic weights and time windows, as presented in an IJCAI competition. The results show that our approach outperforms vanilla ALNS, ALNS tuned with Bayesian optimization, and two state-of-the-art DRL approaches that were the winning methods of the competition, achieving this with significantly fewer training observations. Furthermore, we demonstrate several good properties of the proposed DR-ALNS method: it is easily adapted to solve different routing problems, its learned policies perform consistently well across various instance sizes, and these policies can be directly applied to different problem variants.

show abstract

A survey of adaptive large neighborhood search algorithms and applications

Cited by 29 publications

References 269 publications

Item Storage Assignment Problem in Robotic Mobile Fulfillment Systems With Nonempty Pods

Item Storage Assignment Problem in Robotic Mobile Fulfillment Systems With Nonempty Pods

Adaptive Anytime Multi-Agent Path Finding Using Bandit-Based Large Neighborhood Search

Online Control of Adaptive Large Neighborhood Search Using Deep Reinforcement Learning

Contact Info

Product

Resources

About