Monte Carlo tree search-based meta-heuristic method for faster optimal cluster allocation
Through the branch delimiting and Monte Carlo tree search method combined with the dual graph method, the cost minimization problem of selecting a suitable AMR machine group in the automation facility is solved, and the global optimal machine group composition and task allocation are found within a reasonable time, which improves the material Stream throughput and reduce operational costs.
Patent Information
- Application Number
- CN202410661911.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-06-02
- Filing Date
- 2024-05-27
- Publication Date
- 2024-12-03
AI Technical Summary
In automation facilities, selecting the appropriate autonomous mobile robot (AMR) fleet to minimize costs is a large-scale optimization problem, and prior art is difficult to find global or near-optimal fleet composition and tasks within reasonable computing time. Allocation plan.
The waypoint sequence is determined by branch bounding (B&B) method, the path is determined by the dual graph method, and the group composition and task allocation are determined by combining the Monte Carlo tree search (MCTS) method. Multiple processors are used to divide the search space and randomly walk Sub-update operation costs to enable rapid exploration and local search.
Significantly reduces computing time, finds global or near-optimal cluster composition and task allocation schemes, improves material flow throughput, and reduces operating costs.
Smart Images

Figure CN119067204A_ABST
