Business Analytics Seminar (PhD seminar): Jakob Geertsen Kristensen, AU

Presenter: Jakob Geertsen Kristensen 

Title: Markov Decision Processes for Crop Rotation Problems

Abstract: Crop rotation planning is often based on fixed-rotation policies, in which a predetermined sequence of crops is repeated over time. Such policies are simple to implement, but they may be unable to respond effectively to evolving field conditions, disease pressure, or changing trade-offs between economic and environmental objectives. In this presentation, I examine the potential advantages of dynamic policies for the Crop Rotation Problem (CRP) under multiple evaluation criteria. Markov Decision Processes (MDPs) provide a natural framework for formulating and optimizing such policies, as they account for how current crop choices influence future production conditions and decisions. To illustrate this approach, I present an MDP formulation of a simplified crop-rotation problem involving potatoes and wheat, with objective functions capturing profit, environmental impacts, and disease-related considerations.

Main supervisor: Lars Relund Nielsen

Discussant: Christian Larsen

Organisers: Surabhi Verma and Hartanto Wong