Distributed Chance-Constrained Model Predictive Control for Condition-Based Maintenance Planning for Railway Infrastructures

摘要: We develop a Model Predictive Control (MPC) approach for condition-based maintenance planning under uncertainty railway infrastructure systems composed of multiple components. Piecewise-affine models with uncertain parameters are used to capture both the nonlinearity and uncertainties in deterioration process. To keep balance between robustness optimality, we formulate MPC optimization problem as chance-constrained problem, which ensures that constraints, e.g., bounds on degradation level, satisfied given probabilistic guarantee. Two distributed algorithms, one based Dantzig-Wolfe decomposition other derived from constraint-tightening technique, proposed improve scalability approach. Computational experiments show method performs best terms computational time convergence global optimality. By comparing approaches deterministic approach, traditional time-based despite their high requirements, cost-efficient robust presence uncertainties.