Markov Decision Processes for Services Opportunity Pipeline Optimization

The dynamics of sales opportunities can be modelled by a Markov Decision Process. The latter can be solved using dynamic programming and assigns to each state an optimal action. In this project, states are modelled by the number of opportunities at five different maturity levels called ranks, actions are represented by investments and rewards by profits from signed contracts. Transitions are simulated using the probabilities that an opportunity moves from one rank to another. Two different types of policy appear recurrently in the model outcome, i.e. a low-investment policy when the opportunities are rather uniformly distributed across ranks and a high-investment policy, when a larger number of opportunities have reached mature status or have just entered the pipe. Acknowledgements I would like to thank Dr. Gianluca Antonini for his help with this project and for making corrections to the present report, and Stefan Woerner for his help on the modelling part.