Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Dynamic Team Orienteering Problem in Spatial Crowdsourcing: A Scenario Sampling Approach

Zhibin Wu, Songhao Shen, Yufeng Zhou, Qin Lei

TL;DR

The paper addresses dynamic, profit-maximizing routing for spatial crowdsourcing with heterogeneous worker OD trips and online task arrivals. It proposes Scen-RH-ALNS, a scenario-based, rolling-horizon ALNS framework that augments each planning epoch with virtual tasks to incorporate lookahead and aggregates robust near-term actions across scenarios for joint dispatch. A static HT-TOPTW MIP serves as an offline reference, and extensive experiments on map-based synthetic instances and the DTOP benchmark demonstrate competitive online performance with substantially lower runtimes than multi-plan approaches, plus consistent gains from lookahead. The method offers a scalable, real-time solution for dynamic crowdsourcing tasks and lays groundwork for forecasting integration and stochastic extensions in future work.

Abstract

In services such as retail audits and urban infrastructure monitoring, a platform dispatches rewarded, location-based micro-tasks to mobile workers traveling along personal origin-destination (OD) trips under hard time budgets. As requests with time constraints arrive online over a finite horizon, the platform must decide which requests to accept and how to route workers to maximize collected profit. We model this setting as the Dynamic Team Orienteering Problem in Spatial Crowdsourcing (DTOP-SC). To solve this problem, we propose a scenario-sampling rolling-horizon framework that mitigates myopic bias by augmenting each planning epoch with sampled virtual tasks. At each epoch, the augmented task set defines a deterministic static subproblem solved via an adaptive large neighborhood search (ALNS). We also formulate a mixed-integer programming model to provide offline reference solutions. Computational experiments are conducted on synthetic DTOP-SC instances generated from real-world road-map coordinates and on a dynamic team orienteering (DTOP) benchmark. On the map-based instances, the proposed policy exhibits stable gaps with respect to time-limited MIP solutions across the tested scales, while maintaining smooth computational scalability as the problem size increases. On the DTOP benchmark, the policy achieves an average decision time of 0.14s per instance, with 192-198s reported for multiple plan approach as an indicative reference, while maintaining competitive profit.

The Dynamic Team Orienteering Problem in Spatial Crowdsourcing: A Scenario Sampling Approach

TL;DR

The paper addresses dynamic, profit-maximizing routing for spatial crowdsourcing with heterogeneous worker OD trips and online task arrivals. It proposes Scen-RH-ALNS, a scenario-based, rolling-horizon ALNS framework that augments each planning epoch with virtual tasks to incorporate lookahead and aggregates robust near-term actions across scenarios for joint dispatch. A static HT-TOPTW MIP serves as an offline reference, and extensive experiments on map-based synthetic instances and the DTOP benchmark demonstrate competitive online performance with substantially lower runtimes than multi-plan approaches, plus consistent gains from lookahead. The method offers a scalable, real-time solution for dynamic crowdsourcing tasks and lays groundwork for forecasting integration and stochastic extensions in future work.

Abstract

In services such as retail audits and urban infrastructure monitoring, a platform dispatches rewarded, location-based micro-tasks to mobile workers traveling along personal origin-destination (OD) trips under hard time budgets. As requests with time constraints arrive online over a finite horizon, the platform must decide which requests to accept and how to route workers to maximize collected profit. We model this setting as the Dynamic Team Orienteering Problem in Spatial Crowdsourcing (DTOP-SC). To solve this problem, we propose a scenario-sampling rolling-horizon framework that mitigates myopic bias by augmenting each planning epoch with sampled virtual tasks. At each epoch, the augmented task set defines a deterministic static subproblem solved via an adaptive large neighborhood search (ALNS). We also formulate a mixed-integer programming model to provide offline reference solutions. Computational experiments are conducted on synthetic DTOP-SC instances generated from real-world road-map coordinates and on a dynamic team orienteering (DTOP) benchmark. On the map-based instances, the proposed policy exhibits stable gaps with respect to time-limited MIP solutions across the tested scales, while maintaining smooth computational scalability as the problem size increases. On the DTOP benchmark, the policy achieves an average decision time of 0.14s per instance, with 192-198s reported for multiple plan approach as an indicative reference, while maintaining competitive profit.
Paper Structure (43 sections, 37 equations, 6 figures, 46 tables, 1 algorithm)

This paper contains 43 sections, 37 equations, 6 figures, 46 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related work
  3. Team orienteering problems
  4. Dynamic routing methods
  5. Problem formulation
  6. Operational setting
  7. Dynamic team orienteering problem in spatial crowdsourcing
  8. Static HT-TOPTW snapshot and MIP formulation
  9. Objective function
  10. Constraints
  11. Problem-solving methodology
  12. A rolling-horizon ALNS for DTOP-SC
  13. Initial snapshot construction and static optimization
  14. Event-driven time update and myopic dispatch
  15. Scenario-based extension and multi-worker joint dispatch
  16. ...and 28 more sections

Figures (6)

  • Figure 1: Schematic illustration of the Dynamic Team Orienteering Problem in Spatial Crowdsourcing.
  • Figure 2: Event-driven rolling-horizon decision process in Scen-RH-ALNS.
  • Figure 3: Example of scenario-specific worker-task extraction.
  • Figure 4: Spatial distribution of sampled road-network coordinates used to generate the synthetic instances.
  • Figure 5: Per-instance distributions of $\text{Gap}_{MIP}$ for map-based stability experiments.
  • ...and 1 more figures