Abstract
This paper focuses on the performance comparison of several approximate dynamic programming (ADP) techniques. In particular, we evaluate three ADP techniques through a class of dynamic stochastic scheduling problems: Lagrangian-based ADP, linear programming-based ADP, and direct search-based ADP. We uniquely implement the direct search-based ADP through basis functions that differ from those used in the relevant literature. The class of scheduling problems has the property that jobs arriving dynamically and stochastically must be scheduled to days in advance. Numerical results reveal that the direct search-based ADP outperforms others in the majority of problem sets generated.
Highlights
Approximate dynamic programming (ADP) is a method to solve large-scale Markov decision processes (MDPs), which are used to model systems that evolve stochastically over time
For each problem set determined by the combination of I,C1, and N, columns 2 to 4 of each table give the average discounted cost values over 10 independent problem instances obtained for the Lagrangian-based, the linear programming (LP)-based, and the direct search-based approximate dynamic programming (ADP), respectively
When the discount factor was set to a low level (i.e., 0.9), the direct searchbased ADP outperforms others in all problem sets. (Paired t-tests revealed that the respective percentage differences were statistically significant at the 0.05 level.)
Summary
Approximate dynamic programming (ADP) is a method to solve large-scale Markov decision processes (MDPs), which are used to model systems that evolve stochastically over time. The term approximate refers to the fact that the solution obtained by the underlying ADP technique is an approximate to the optimal solution. ADPs have been used to solve problems arising in diverse fields such as healthcare, manufacturing, transportation, and revenue management. In the last few decades, various ADP techniques have been proposed to approximately solve computationally intractable MDPs. The state-of-theart ADP techniques include the Lagrangian-based ADP ( [1], [2]), the linear programming-based ADP ( [3], [4]), and the direct search-based ADP techniques ( [5], [6]). The performances of those techniques have not been evaluated in the literature
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: An International Journal of Optimization and Control: Theories & Applications (IJOCTA)
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
