Source-linked AI summary
Collaborative vehicle routing: a survey
Margaretha Gansterer, Richard F. Hartl
TL;DR
Collaborative vehicle routing addresses how carriers can jointly plan logistics operations to improve efficiency and sustainability. The survey structures the literature across centralized and two decentralized planning streams, reviews their methods and findings, and identifies open research needs. Reported centralized collaboration benefits commonly reach 20–30%, while uncertainty and more complex transportation systems remain insufficiently studied.
Problem
The literature on collaborative vehicle routing is extensive but requires a structured review spanning centralized planning, decentralized settings, and recent studies.
Method
The survey classifies collaborative vehicle-routing studies by planning problem and collaboration setting across centralized, non-auction-based decentralized, and auction-based decentralized planning.
Results
Centralized collaborative planning studies report potential benefits commonly around 20–30% relative to non-cooperative solutions, including a 25.6% travel-distance improvement in city logistics.
Takeaways & Limitations
The survey provides a structured overview of collaborative vehicle-routing research and identifies centralized, non-auction-based decentralized, and auction-based decentralized planning as its three major streams.
Takeaways & Limitations
Evidence on collaboration under uncertainty is scarce, and gains in complex multimodal or multidepot systems remain insufficiently investigated.
Abstract
from arXiv · showhide
In horizontal collaborations, carriers form coalitions in order to perform parts of their logistics operations jointly. By exchanging transportation requests among each other, they can operate more efficiently and in a more sustainable way. Collaborative vehicle routing has been extensively discussed in the literature. We identify three major streams of research: (i) centralized collaborative planning, (ii) decentralized planning without auctions, and (ii) auction-based decentralized planning. For each of them we give a structured overview on the state of knowledge and discuss future research directions.
1. Introduction
Collaborative vehicle routing is presented as a practically important response to competitive pressure and environmental concerns. The survey addresses gaps in earlier reviews by covering centralized planning, recent literature, and a broader classification.
- Collaborative vehicle routing covers cooperation intended to increase vehicle-fleet operating efficiency while also serving ecological goals.Collaborations are linked to reduced emissions, congestion, and noise pollution, as well as improved service levels and capacities.
- Earlier reviews omitted either collaborating shippers, centralized planning, or operational planning problems, leaving important literature uncovered.About 45% of related literature concerns centralized planning, while nearly 60% of related articles were published in the last three years.
- The survey contributes by including centralized collaborative planning, reviewing recent studies, and introducing a broader article classification.
- The survey is organized around methodology, classifications and definitions, centralized planning, decentralized planning with and without auctions, future research, and a conclusion.
2. Research methodology
The survey focuses on operations-research models and solution techniques for collaborative transportation planning. It identifies relevant journal studies through database searches, reference screening, and citation-based expansion.
- The review includes studies applying operations-research models and solution techniques to operational planning problems.Pure empirical studies and studies focused mainly on coalition design without transportation planning are excluded.
- Relevant studies were identified by combining collaboration-related and transportation-related search terms across library databases.The terms covered collaboration, cooperation, coalition, alliance, transportation, routing, logistics, freight, carrier, and shipper.
- The authors restricted the initial review to journal publications to survey a reasonable number of articles.
- The search was expanded by screening reference lists and then articles citing the relevant studies.The citation-based descendancy approach proceeded from older information toward newer information.
3. Classifications and definitions
The survey distinguishes collaborative vehicle-routing settings by planning structure, routing formulation, shipment type, participants, solution methodology, and benefit sharing. Centralized planning jointly maximizes total profit, whereas decentralized planning exchanges selected requests with limited information and may use auctions.
- Planning structure: Centralized planning jointly maximizes total profit, while decentralized planning exchanges subsets of requests while revealing no or only limited information.Decentralized studies are further divided into non-auction-based and auction-based planning.
- Routing formulation: Routing studies are classified as VRP, ARP, IRP, LCP, MCFP, and assignment problems according to their modeled operational structure.The categories cover node-based routing, arc traversal, inventory integration, lane coverage, network flows, and vehicle-to-request assignment.
- Model extensions: Models may include delivery time windows and pickup-and-delivery requests, where pickup and delivery locations need not coincide with a depot.These extensions capture timing constraints and requests involving separate pickup and delivery nodes.
- Shipment type: The literature distinguishes FTL and LTL shipments, with FTL treated as a special case of LTL when order size equals vehicle capacity.LTL commonly concerns multiple small-volume products, whereas FTL often transports a single product.
- Collaboration participants: Collaborating players may be carriers or shippers, although joint routing planning is typically assigned to carriers and decentralized settings must account for information asymmetries.Carriers own and operate transportation equipment, while shippers own or supply shipments.
- Solution and benefit sharing: Surveyed studies use both exact and heuristic solution methodologies, while benefit sharing commonly draws on the Shapley value, proportional methods, or the nucleolus.The survey also seeks to connect transportation-problem research with profit-sharing research, which are rarely combined in the same studies.
4. Centralized collaborative planning
Centralized collaborative planning studies collaboration under full information, assessing gains against non-cooperative settings and developing models and solution approaches. Reported benefits commonly reach 20–30%, while ecological gains and methodological challenges motivate further research.
- Research streams: Centralized collaborative planning assumes a central authority with full information and includes both collaboration-gain assessment and methodological contributions.The surveyed studies address collaborative planning beyond simply solving a standard optimization problem, including decisions about outsourcing and request exchange.
- Collaboration gain assessment: Joint route planning can achieve synergy values of up to 30% compared with independently optimized routes.Cruijssen et al. examine multiple companies with separate distribution orders under cooperative and non-cooperative planning.
- Collaboration gain assessment: Many studies report total-profit improvements of around 20–30% over non-cooperative solutions, with coalition benefits marginally decreasing as partnership size grows.A real-world courier application also reports up to 20% lower travel cost under cooperative routing.
- Collaboration gain assessment: Collaborative routing can improve ecological outcomes alongside economic performance, including 25.6% lower travel distance, 60% lower greenhouse-gas emissions, and nearly 55% cost savings.Other experiments report cost benefits of about 4–24% and aggregated emission benefits of about 8–33%.
- Methodological contributions: Methodological contributions address interconnected routing and assignment decisions through decomposition, integer programming, branch-and-cut, Lagrangian relaxation, Benders decomposition, and models combining horizontal and vertical cooperation.The surveyed problems extend beyond standard VRP formulations to include ARP, MCFP, and pickup-and-delivery settings.
- Future research directions: Future research should examine uncertain and more complex transportation systems, information-revelation incentives, and methods balancing total cost against deviation from decentralized solutions.The survey notes that most collaboration-gain studies assume deterministic scenarios and that central authorities face highly complex optimization problems.
5. Decentralized planning without auctions
Decentralized non-auction-based planning addresses partner selection, request selection, and request exchange when collaborators reveal no or limited information. These approaches reduce procedural complexity but provide less information about collaborators’ preferences, limiting collaboration profits.
- Decentralized planning is divided into partner selection, request selection, and request exchange.
- Request Selection: Carriers select requests to offer while retaining some requests for their private fleets, and coordinating authorities seek feasible assignments to carriers.
- Studies address capacity acquisition, empty-backhaul reduction, vertical collaboration, lane exchanges, and decision support for dynamic routing.
- Request Selection: 27% cost savings are reported in a real-world freight-data analysis of optimization models and heuristics for adding partners’ pickup-and-delivery tasks.
- Request Exchange: Collaborative request exchange can use route-based mechanisms in which carriers iteratively generate and submit routes using feedback from an agent.
- Future research directions: Non-auction-based systems are generally less complex than auctions, but limited preference information can lead to relatively low collaboration profits.
6. Auction-based decentralized planning
Auction-based decentralized planning exchanges requests through a central auctioneer, using bids and bundle allocation to coordinate carriers with limited information sharing. The literature develops combinatorial and multi-round mechanisms, but integrated, practically usable auction frameworks remain incomplete.
- Auctions let carriers act as both buyers and sellers while an auctioneer coordinates decentralized request exchange.
- Combinatorial auctions bundle requests because individual requests may be unattractive unless combined with others, and accepted bids transfer the full bundle.
- The auction process comprises request pooling, bundle generation, bidding, winner determination, and profit sharing.
- Request selection: Request selection must identify requests valuable to collaborators without requiring carriers to reveal sensitive information, with geographical criteria outperforming pure profit-based strategies.
- Multi-round auctions reduce complexity by offering subsets of items over successive rounds and reusing previously gained information.
- Future research directions: Auctions may increase collaboration profits, but the five auction phases each contain complex, partly unsolved decisions and their interactions remain insufficiently investigated.
- Future research directions: Strategic behavior in request selection and bidding, effective profit sharing, auction comparisons, and the value of information remain open research directions.
7. Conclusion
The survey classifies collaborative vehicle-routing research into centralized, non-auction-based decentralized, and auction-based decentralized planning. It identifies integration, strategic behavior, comparison, information value, and benchmarking as priorities for future work.
- The survey provides a structured classification of collaborative vehicle-routing literature across three major planning streams.
- Future work includes applying collaborative frameworks to more complex transportation systems, including multimodal settings.
- Researchers should investigate strategic behavior and profit-sharing mechanisms intended to avoid it.
- Comparing auction-based with non-auction-based systems and assessing the value of information are identified as open directions.
- An open repository of benchmark instances would support comparable evaluation of performance gaps between centralized and decentralized approaches.