An optimal route selection with reinforcement learning for the waste management in Dhaka city

Abstract

In the capital city of Dhaka, currently, one of the biggest problems is its waste management. Because of over-population, the amount of waste produced every day is so high that it can not be handled properly with the infrastructural aid available now. Transporting the wastes faces a lot of obstacles because of the road and the traffic conditions. As a result, the whole operation is delayed. Furthermore, maintaining a hygienic situation becomes an impossible task. In this work, a modified system will be proposed for dumping wastes in different parts of the city. By using two reinforcement learning techniques (Q-learning, SARSA) imposed in this model, the system will be allowed to find the optimal route for the waste-carrying vehicle so that a faster transportation is ensured so that a suitable state of the environment can be sustained.