MULTIPATH ROUTING IN IoT USING REINFORCEMENT LEARNING
Wireless Sensor Networks is generally described as a large network of nodes for data monitoring, collection and dissemination. A Large number of routing protocols have been used to achieve actual output and reliability and also involved many numbers of nodes to obtain enormous count of paths which therefore reduces the reliability ofIoT services.Here, a multipath routing protocol is introduced using reinforcement learning.Using this reinforcement method best possible paths from source to destination can be found. The algorithm we use here is Q-learning. Advantage of using this technique is that there is no necessary for this method to depend on environments and they can find the solution for the problem using rewards and transitions. This algorithm tells the representative what action should be taken under what situations. Q-learning algorithm finds the solution to the problem for choosing the efficient path from the source node to the destination node in very less time than the path bridging technique. It is designed to explore possible routes and adapts a network routing based on information gathered and converges towards an efficient solution.