Decentralized Route Choice Control Of Automated Baggage Handling Systems.pdf

This technical report by Delft University of Technology's Delft Center for Systems and Control presents a decentralized approach to route choice control in automated baggage handling systems using destination coded vehicles (DCVs). The paper addresses the challenge of optimizing bag transportation through a network of conveyors and junctions, aiming to arrive at end points within specified time windows.

System Description:
- DCV-based system with scanners, screening equipment, conveyors, and DCVs transporting bags.
- Higher-level control problems: route assignment for bags and line balancing for empty DCVs.
- Low-level control: velocity control of DCVs, coordination during loading/unloading, and switch control at junctions.

Event-Driven Model:
An event-driven model simulates the system, including continuous movement of DCVs and discrete events like bag loading/unloading, switch updates, and speed adjustments.

Proposed Control Methods:

1. Decentralized Model Predictive Control (MPC):
- Centrally computes optimal routes for each DCV considering future events.
- Decentralized implementation requires each DCV to communicate local information and receive global constraints.
- Offers near-optimal solutions but incurs significant computational overhead.

2. Decentralized Heuristic Approaches:
- Two heuristics are proposed:
- Fast Heuristic: Rapidly calculates routes based on simple rules, prioritizing direct paths.
- Adaptive Heuristic: Adjusts routes dynamically based on current system state and traffic conditions.
- Require minimal computation time but generally yield suboptimal results compared to MPC.

Comparison and Analysis:
- Simulation results demonstrate the performance of decentralized MPC and heuristic approaches.
- Decentralized MPC provides near-optimal solutions, while heuristics offer faster but less efficient routes.
- Heuristics are advantageous for real-time control due to their low computational complexity.

Conclusions:
The paper concludes that decentralized MPC offers high-quality route planning, while decentralized heuristics provide efficient, real-time alternatives for automated baggage handling systems.