Disaster response and resource allocation during natural or man-made disasters are critical but highly challenging tasks. The problem at hand is the allocation of resources in humanitarian logistics. It deals majorly with allocating resources in as efficient a manner as possible.The problem deals with multiple areas (affected regions) that require resource supplies from resource centers (single or multiple), and decisions are made about how to distribute resources to minimize costs while considering factors like accessibility, deprivation, and penalties.
We have compared the Q-learning and Deep Q-learning algorithms to check how they differ in rewards and time taken to solve the problem of resource allocation.The capacity of resource centers can be multiple units.The demand of affected area in each time period can also vary from one-unit to multiple units.There are total three cases considered. One is allocation of resources by single resource center to multiple affected areas. Second is allocation of resources from multiple resource centers to multiple affected areas in such a way that at a single time period, only single demand of affected are could be fulfilled. Third one is for multiple resource centers serving multiple affected areas such that in one time period, multiple demands of affected areas could be fulfilled.