Comparing reinforcement learning approaches for solving game theoretic models: a dynamic airline pricing game example

Download

HTML - Publishers print Restricted to Repository staff only Download (89Kb) | Request a copy

Description/Abstract

Due to the difficulty in solving game theoretic models, there is a tendency to focus on the overly simplistic dynamic airline pricing games or to even ignore competition completely. Recent changes to the industry mean that airlines can no longer ignore competitors in their model. This paper adds more complex customer model aspects; i.e., customer choice using a Logit model, customer demand using a linear probabilistic demand model, and market size using a binary random function; into an existing solvable game that only had a simple customer model. A reinforcement learning method was used to solve the newly formed games with mixed results.