Implementation of Q-Factor (Knowledge of reinforcement learning is necessary )

Implementation of Q-Factor (Knowledge of reinforcement learning is necessary ) Please read task detail before biding • We have one small network, that has consisted on 5 nodes and one sink (It will collect the data from the whole network through multi-hop communication). • We have already coded some work (In Python 3.0) and work is related to reinforcement learning (Q-learning) which is also included in python code. The code is divided into two threads (Control data which send the control signal and 2nd thread is data (this thread is not active yet)) • I want to add some functions in existing code(That I will provide you), which is running well. • I will also provide the function formulas. Following implementation is required in Python • Data traffic which will be two types. One Constant bitrate (i.e., 1Kb)and constant frequency (i.e., data packet send after each 1s). 2nd is variable bitrate (i.e., 1kb,2kb,3kb,4kb) and variable sending frequency • We divide the system working into two category 1. Request handling (Control signaling like hello message from neighbor nodes) 2. Neighbor selection (where the current node will send the data ) • We divide the request handling into four types 1. Always accepted (This is already implemented and working well) 2. Distance-based accept (Formulas will be provided by me) 3. Load balance aware accept (Formulas will be provided by me) 4. Quality conserving aware accept (Formulas will be provided by me) • We divide the Neighbor selection into three types 1. e-greedy (Already implemented and working well) 2. Adaptive e-greedy (Formulas will be provided by me) 3. e-multi criteria (Formulas will be provided by me) • I will tell the data packet content (Mean which things we need at sink i.e packet number, sending time, remaining battery, etc) which need to send up to sink node. • For whole work, you need at least two to three pcs/raspberries for making ad-hoc network and testing purpose. • The code should be generalized which could implement on all topologies