Optimization and Machine Learning Applied to Last-Mile Logistics: A Review

Sustainability, Journal Year: 2022, Volume and Issue: 14(9), P. 5329 - 5329

Published: April 28, 2022

The growth in e-commerce that our society has faced recent years is changing the view companies have on last-mile logistics, due to its increasing impact whole supply chain. New technologies are raising users’ expectations with need develop customized delivery experiences; moreover, pressure chains also created additional challenges for suppliers. At same time, this phenomenon generates an increase liveability of cities, traffic congestion, occupation public spaces, and environmental acoustic pollution linked urban logistics. In context, optimization deliveries imperative not only parcels be delivered areas, but administrations want guarantee a good quality life citizens. years, many scholars focused study logistics techniques and, particular, last mile. addition traditional techniques, disciplines operations research, advances use sensors IoT, consequent large amount data derives from it, pushing towards greater big analytics techniques—such as machine learning artificial intelligence—which sector. Based premise, aim work provide overview most literature related techniques; used baseline who intend explore new approaches optimization. A bibliometric analysis critical review were conducted order highlight main studied problems, algorithms used, case studies. results allow studies clustered into models, approaches, mixed methods. research gaps limitations current assessed identify unaddressed suggestions future approaches.

Language: Английский

---

Differential Evolution With a Variable Population Size for Deployment Optimization in a UAV-Assisted IoT Data Collection System

IEEE Transactions on Emerging Topics in Computational Intelligence, Journal Year: 2019, Volume and Issue: 4(3), P. 324 - 335

Published: Oct. 1, 2019

This paper studies an unmanned aerial vehicle (UAV)-assisted Internet of Things (IoT) data collection system, where a UAV is employed as platform for group ground IoT devices. Our objective to minimize the energy consumption this system by optimizing UAV's deployment, including number and locations stop points UAV. When using evolutionary algorithms solve deployment problem, each individual usually represents entire deployment. Since unknown priori, length in population should be varied during optimization process. Under condition, variable-length problem traditional fixed-length mutation crossover operators modified. In paper, we propose differential evolution algorithm with variable size, called DEVIPS, location point encoded into individual, thus whole Over course evolution, produce offspring. Afterward, design strategy adjust size according performance improvement. By strategy, can increased, reduced, or kept unchanged adaptively. since has fixed length, becomes used directly. The DEVIPS compared that five on set instances. experimental demonstrate its effectiveness.

Language: Английский

---

An Overview and Experimental Study of Learning-Based Optimization Algorithms for the Vehicle Routing Problem

IEEE/CAA Journal of Automatica Sinica, Journal Year: 2022, Volume and Issue: 9(7), P. 1115 - 1138

Published: June 30, 2022

The vehicle routing problem (VRP) is a typical discrete combinatorial optimization problem, and many models algorithms have been proposed to solve the VRP its variants. Although existing approaches contributed significantly development of this field, these either are limited in size or need manual intervention choosing parameters. To difficulties, studies considered learning-based (LBO) VRP. This paper reviews recent advances field divides relevant into end-to-end step-by-step approaches. We performed statistical analysis reviewed articles from various aspects designed three experiments evaluate performance four representative LBO algorithms. Finally, we conclude applicable types problems for different suggest directions which researchers can improve

Language: Английский

---

Designing efficient algorithms for logistics management: Optimizing time-constrained vehicle routing

Strategic Management, Journal Year: 2025, Volume and Issue: 00, P. 83 - 83

Published: Jan. 1, 2025

Background: City logistics is a critical component of urban economic development, as it optimizes supply chains, enhances customer satisfaction through reliable deliveries, and minimizes environmental impacts in densely populated areas. This field addresses various challenges, including traffic congestion, concerns, noise pollution, the crucial need for timely deliveries. Routing scheduling are central to operations, with modern software integrating time windows meet precise demands driven by detailed requirements operational efficiencies. Furthermore, advanced vehicle routing models now effectively simulate real-world factors such stochastic travel times, dynamic product demands. Purpose: paper aims develop an algorithm that decisions. Our approach extends dimension, considering times service within predefined windows. Study design/methodology/approach: The proposed structured execute iterative phases, aiming optimize key logistical objectives. In order generate competitive solutions, we seek minimize number vehicles utilized overall costs. evaluation solution space was conducted via Simulated Annealing. Findings/conclusions: performance algorithm, evaluated using Gehring Homberger benchmark instances 200 customers, demonstrates its effectiveness. successfully meets target required, associated costs on average 1% best solutions reported relevant literature. Limitations/future research: Given ongoing from decision-makers, future research endeavors will focus enhancing computational efficiency algorithm. Additionally, incorporating more time-related features, could further improve algorithm's real-time applicability.

Language: Английский

---

Two meta-heuristics for solving the capacitated vehicle routing problem: the case of the Tunisian Post Office

Operational Research, Journal Year: 2020, Volume and Issue: 22(1), P. 507 - 549

Published: Jan. 4, 2020

Language: Английский

---

Improved social spider algorithm for partial disassembly line balancing problem considering the energy consumption involved in tool switching

International Journal of Production Research, Journal Year: 2022, Volume and Issue: 61(7), P. 2250 - 2266

Published: July 14, 2022

As the waste products have a variety of connection structure characteristics, energy consumed in tool switching disassembly process is considered to better comprehensively optimise consumption index. A mixed-integer non-linear programming (MINLP) model multi-objective partial line balancing problem (PDLBP) constructed minimise four optimisation objectives which are number workstations, workstation load, tools switched, and consumption. Based on characteristics PDLBP, we an matrix proposed improved social spider algorithm (ISSA). The random movement mask change operations ISSA were improved, artificial spiders added enhance global capabilities ISSA. was applied two typical benchmark instances, different scales, respectively. And computational results compared with several algorithms existing literature verify superiority Finally, instance printer, switching. Then, multiple schemes provided for decision-makers.

Language: Английский

---

The ground handler dock capacitated pickup and delivery problem with time windows: A collaborative framework for air cargo operations

Transportation Research Part E Logistics and Transportation Review, Journal Year: 2022, Volume and Issue: 159, P. 102603 - 102603

Published: Feb. 7, 2022

We study a typical problem within the air cargo supply chain, concerning transportation of standard Unit Load Devices (ULDs) from freight forwarders' to ground handlers' warehouses. First, ULDs are picked up by set available trucks at premises time window. Next, they delivered handlers, also window, and discharged according Last In First Out (LIFO) policy. Due space constraints, handlers have limited capacity serve waiting times may arise, especially in case forwarders do not coordinate their operations. Therefore, this paper we consider cooperative framework where is coordinated central planner. The goal planner find proper routing scheduling that minimizes sum warehouses, while satisfying trucks. propose two mathematical formulations, one based on other packing aspect problem. To solve large instances problem, an Adaptive Large Neighborhood Search algorithm developed. With numerical experiments, compare performances models metaheuristic, quantify benefits proposed reduce times.

Language: Английский

---

Towards Faster Vehicle Routing by Transferring Knowledge From Customer Representation

IEEE Transactions on Intelligent Transportation Systems, Journal Year: 2020, Volume and Issue: 23(2), P. 952 - 965

Published: Sept. 10, 2020

The Vehicle Routing Problem (VRP) is a well-known NP-hard combinatorial optimization problem, which has wide spread applications in real world, such as logistics, bus route planning, and urban path planning. To solve VRP, traditional methods usually start the search from scratch ignore VRPs solved past, could lead to repeated explorations of space related problems, thus results slow process involving unnecessary computational cost. Keeping this mind, speed up for vehicle routing, article presents new study towards faster routing by transferring knowledge customer representations are learned past VRPs. In particular, we propose capture useful traits buried previous optimized solutions learning representation, can be transferred across VRPs, serving prior knowledge, bias target VRP. contrast existing approaches, proposed transfer consist representation based on solution, general possessing different structural properties, weighted $l_{1}$ norm-regularized formulation building sparse mapping that easy solve. Further, occurs along whole process, able guide consistently. verify efficacy method, using population-based method VRP solver, comprehensive empirical studies both commonly used benchmarks world application presented.

Language: Английский

---

An Improved Genetic Algorithm for the Granularity-Based Split Vehicle Routing Problem with Simultaneous Delivery and Pickup

Mathematics, Journal Year: 2023, Volume and Issue: 11(15), P. 3328 - 3328

Published: July 28, 2023

The Split Vehicle Routing Problem with Simultaneous Delivery and Pickup (SVRPSDP) consists of two subproblems, i.e., the (VRPSDP) (SDVRP). Compared to SVRPSDP is much closer reality. However, some realistic factors are still ignored in SVRPSDP. For example, shipments integrated cannot be infinitely subdivided. Hence, this paper investigates Granularity-based (GSVRPSDP). characteristics GSVRPSDP that demands customers split into individual both volume weight each shipment considered. In order solve efficiently, a Genetic-Simulated hybrid algorithm (GA-SA) proposed, which Simulated Annealing (SA) inserted Genetic Algorithm (GA) framework improve global search abilities individuals. experimental results indicate GA-SA can achieve lower total costs routes compared traditional meta-algorithms, such as GA, SA Particle Swarm Optimization (PSO), reduction more than 10%. further analysis, space utilization capacity vehicles calculated, 86.1% 88.9%, respectively. These values higher those achieved by GA (71.2% 74.8%, respectively) PSO (60.9% 65.7%, respectively), confirming effectiveness GA-SA. And superiority simultaneous delivery pickup proved comparing separate pickup. Specifically, 80%

Language: Английский

---

Vehicle routing: Review of benchmark datasets

Journal of the Operational Research Society, Journal Year: 2021, Volume and Issue: 72(8), P. 1794 - 1807

Published: Feb. 19, 2021

Language: Английский

---