Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Repair Crew Routing for Infrastructure Network Restoration under Incomplete Information

Subhojit Biswas, Bahar Cavdar, Joseph Geunes

TL;DR

The paper addresses repair crew routing for infrastructure restoration under incomplete information, formulating the Traveling Repairman Network Restoration Problem (TRNRP) as a finite-horizon Markov decision process. It advances solution methods by introducing a reinforcement-learning-based approximate dynamic programming framework that uses post-decision states and a problem-specific state transformation to manage information revelation. Structural results prune suboptimal routes and three state-aggregation schemes drastically reduce the search space, enabling larger-scale instances to be tackled. Computational experiments show that the proposed NR R and its strengthened sNRR variants outperform benchmark heuristics across multiple network sizes, with substantial reductions in training time and memory when using aggregation. The work enables scalable, near-optimal routing for power restoration and suggests future extensions to multiple crews and dynamic network-switching decisions.

Abstract

This paper considers a disrupted infrastructure network where the repair crew knows the locations of service outages but not the locations of actual faults. Our goal is to determine a route for a single crew to visit and repair the disruptions to restore service with minimum negative impact. We call this problem the Traveling Repairman Network Restoration Problem (TRNRP). This problem presents strong computational challenges due to the combinatorial nature of the decisions, inter-dependencies within the underlying infrastructure network, and incomplete information. Considering the dynamic nature of the decisions as a result of dynamic information revelation on the status of the nodes, we model this problem as a finite-horizon Markov decision process. Our solution approach uses value approximation based on reinforcement learning, which is strengthened by structural results that identify a set of suboptimal moves. In addition, we propose state aggregation methods to reduce the size of the state space. We perform extensive computational studies to characterize the performance of our solution methods under different parameter settings and to compare them with benchmark solution approaches.

Repair Crew Routing for Infrastructure Network Restoration under Incomplete Information

TL;DR

The paper addresses repair crew routing for infrastructure restoration under incomplete information, formulating the Traveling Repairman Network Restoration Problem (TRNRP) as a finite-horizon Markov decision process. It advances solution methods by introducing a reinforcement-learning-based approximate dynamic programming framework that uses post-decision states and a problem-specific state transformation to manage information revelation. Structural results prune suboptimal routes and three state-aggregation schemes drastically reduce the search space, enabling larger-scale instances to be tackled. Computational experiments show that the proposed NR R and its strengthened sNRR variants outperform benchmark heuristics across multiple network sizes, with substantial reductions in training time and memory when using aggregation. The work enables scalable, near-optimal routing for power restoration and suggests future extensions to multiple crews and dynamic network-switching decisions.

Abstract

This paper considers a disrupted infrastructure network where the repair crew knows the locations of service outages but not the locations of actual faults. Our goal is to determine a route for a single crew to visit and repair the disruptions to restore service with minimum negative impact. We call this problem the Traveling Repairman Network Restoration Problem (TRNRP). This problem presents strong computational challenges due to the combinatorial nature of the decisions, inter-dependencies within the underlying infrastructure network, and incomplete information. Considering the dynamic nature of the decisions as a result of dynamic information revelation on the status of the nodes, we model this problem as a finite-horizon Markov decision process. Our solution approach uses value approximation based on reinforcement learning, which is strengthened by structural results that identify a set of suboptimal moves. In addition, we propose state aggregation methods to reduce the size of the state space. We perform extensive computational studies to characterize the performance of our solution methods under different parameter settings and to compare them with benchmark solution approaches.
Paper Structure (15 sections, 2 theorems, 13 equations, 10 figures, 11 tables, 5 algorithms)

This paper contains 15 sections, 2 theorems, 13 equations, 10 figures, 11 tables, 5 algorithms.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Problem Definition
  4. Formulation
  5. Solution Approach
  6. Refinements
  7. State Aggregation
  8. Computational Experiments
  9. Conclusion
  10. Algorithms
  11. Flow Charts
  12. Proof of Proposition \ref{['Prop:PrereqRefinement']}
  13. Proof of Proposition 2
  14. Instance Generation
  15. Extended Computational Results

Key Result

Proposition 1

Consider nodes ${i, j , k} \in N_{\mathcal{I}}$ such that $k \preceq j$ on the infrastructure network. If $d_{ij} \leq d_{ik}$ and $d_{kh} \leq d_{jh}$ and node $h$ is not the depot, or $d_{ij} \leq d_{ik}$ and node $h$ is the depot, then any offline solution to the crew routing problem that has $i\

Figures (10)

  • Figure 1: A representation of a distribution network (IEEE 13-node instance).
  • Figure 2: Possible node transitions for $a \in U^{-}_p$ (left) and for $a \in U^{-}_1$ (right).
  • Figure 3: Node dispersions of power distribution networks with 20 nodes on a $10 \times 10$ square region.
  • Figure 4: Node dispersions of power distribution networks with 30 nodes on a 10$\times$10 square region.
  • Figure 5: Node dispersions of power distribution networks with 50 nodes on a 10$\times$10 square region.
  • ...and 5 more figures

Theorems & Definitions (7)

  • Proposition 1
  • Definition 1
  • Definition 2
  • Definition 3
  • Proposition 2
  • proof
  • proof