Integrated condition-based maintenance and multi-item lot-sizing with stochastic demand

This paper studies the problem of integrated lot-sizing and maintenance decision making in case multiple products stochastic demand. The is formulated as a Markov process, which goal to find joint production policy that minimizes long run expected total discounted cost. Therefore, classic Q-learning algorithm adopted, decomposition-based approximate Q-value heuristic developed obtain near-optimal solutions reasonable time. To accelerate convergence algorithm, hybrid method proposed Q-values are initiated by output heuristic. numerical experiments reveal outperformed algorithms terms accuracy converges much faster than method. However, these so-called tabular methods do not scale larger problems with more four products. Hence, based on structure, three state aggregation schemes applied solve large-scale problems. study demonstrates third scheme performs nearly good while significantly reducing computational time being scalable