Order Acceptance in dynamic multi-project environment with stochastic renewable resource using Dynamic Programming approach

Order Acceptance has been discussed in studies in a wide variety of environments. Importance of selecting projects among project portfolio for operation in a scarce resource environment is not obscure for anyone. In this study, we consider working on stochastic arrival of projects to consider a dynamic environment with stochastic project durations and constrained resource. The amount of resource in each decision epoch is stochastic. We are looking for an optimal policy for the acceptance of projects that their reward is known as well as is specific for each project. We develop Linear Program of Markov Decision Process and examine it with an example. To solve the example, we employ Linear Programming, modified policy iteration, and Q-learning using MATLAB software.

Keywords:

Order Acceptance , constrained resource , stochastic project arrival , Markov Decision Process