Policy Optimisation for the drug/vaccine disribution problem.
Description of Problem Statement
- In the event of a suitable drug/vaccine hitting the market, we need an optimal policy to decide the basis for its distribution. The number of variables that then come into play in this decision making process are huge, and hence an attempt has been made to use an AI framework as the decision-support system for the problem statement in question. Reinforcement learning has been employed for this task.
- There are certain differences that arise during the distribution of drugs and vaccines, owing to the difference in their action(curative/preventive) against
the Sars-Cov2 virus. These differences have been identified and analysed to be reproduced into two different environments, for drug (
gym-drug) and vaccine (gym-vaccine) respectively.