Dinamik stokastik ileri programlama için doğrudanarama tabanlı yaklaşımlı dinamik programlamatekniği

Dynamic stochastic scheduling problems are one of the most critical and challenging problems in optimization. In this thesis, we study dynamic stochastic scheduling problems with cancellations. In these problems, jobs arrive randomly at a system and have deadlines. We present a strategy for solving this problem, which involves formulating the problem through Markov Decision Process (MDP), and then solving it approximately using a direct search-based Approximate Dynamic Programming (ADP) technique. We perform the performance comparison of the direct search-based ADP policy and the Greedy policy under diverse scenarios. Our numerical results reveal that the greedy policy can be significantly improved through the implementation of the direct search-based ADP. Keywords: Dynamic scheduling, Markov decision processes, Dynamic programming,