Assignment Instructions

Develop a SARSA agent that$$is capable of acquiring$$policies within the$$openAI$$Gym Frozen Lake setting.$$Frozen Lake presents a grid world scenario in which the agent$$is required to$$navigate a treacherous frozen surface with perilous holes that could lead to fatal consequences.$$Procedure$*$ Instantiate the frozen Lake environment with gym. make $($Frozen Lake$-$V$1,$desc$=$desc$)$ and use unwrapped method.$*$ The input seed should be used to seed the random number generators for both Gym and$$Numpy$.*$ Implement SARSA agent using an epsilon greedy$$behavioral$$policy$.*$ Initialize the agents Q table to zeros$*$ Train your agent with the given parameters. The input$$amap$$is the Frozen Lake map that you need to resize and provide to the$$desc$$attribute when you instantiate the environment, gamma is the discount rate, alpha is the learning rate, epsilon is the parameter for epsilon greedy$$behavior$$strategy your$$sarsa$$agent will use. An action should be selected uniformly at random if a random number is drawn between $0$ and $1$ is less than epsilon.$*$ Gamma$=1\mathrm{.}0,$ Alpha$=0\mathrm{.}25,$ Epsilon$=0\mathrm{.}29,$ No$\_$of$\_$episodes$=14697,$ seed$=741684,$amap$=["$SFFFHFFFFFFFFFFG$"]$