622. A Self-Adaptive Monte Carlo Tree Search Algorithm for Generalized Quay Crane Scheduling Problem
Invited abstract in session MB-32: Seaside Planning-2, stream Maritime and Port Logistics.
Monday, 10:30-12:00Room: Maurice Keyworth 1.09
Authors (first author is the speaker)
| 1. | Chenhao Zhou
|
| School of Management, Northwestern Polytechnical University | |
| 2. | Lei Hai
|
| Northwestern Polytechnical University | |
| 3. | Li Xue
|
| Northwestern Polytechnical University |
Abstract
Efficient quay crane (QC) handling is crucial for enhancing service levels and competitiveness in container terminals, particularly with the advent of ultra-large vessels necessitating rapid container turnover. As the complexity of terminal operations escalates, this paper addresses the generalized quay crane scheduling problem (GQCSP) aiming to minimize makespan for discharging and loading operations, where QCs operate bi-directionally under safety distance and non-crossing constraints, emphasizing the solution approach that delivers swift, feasible solutions over protracted optimal ones, essential for adapting to last-minute operational changes. The problem is formulated as a Markov decision process model, and a self-adaptive Monte Carlo tree search algorithm (MCTS) is proposed that can handle up to 10 QCs and 280 container groups. To expedite the MCTS, problem-specific methods are developed, achieving an improvement of approximately 1.2% of the objective values and a reduction of two-thirds of the computation time. Specifically, an adaptive lower bound is introduced for node selection and subtree pruning, accompanied by a lower bound-based reward function to facilitate backpropagation.
This research showcases the MCTS algorithm's capability to significantly outperform existing heuristics, improving operational efficiency by an average of 20.92% to 72.88% across various scenarios, and achieving near-optimal solutions with remarkable speed.
Keywords
- Maritime applications
- Scheduling
Status: accepted
Back to the list of papers