kayzwer / generalized-policy-iteration-research Goto Github PK

 GridWorld gridWorld = new GridWorld(4, 4, 1.0);
 gridWorld.state_rewards[0][1] = -1.0;
 gridWorld.state_rewards[0][2] = -1.0;
 gridWorld.state_rewards[0][3] = -1.0;
 gridWorld.state_rewards[1][0] = -1.0;
 gridWorld.state_rewards[1][1] = -1.0;
 gridWorld.state_rewards[1][2] = -1.0;
 gridWorld.state_rewards[1][3] = -1.0;
 gridWorld.state_rewards[2][0] = -1.0;
 gridWorld.state_rewards[2][1] = -1.0;
 gridWorld.state_rewards[2][2] = -1.0;
 gridWorld.state_rewards[2][3] = -1.0;
 gridWorld.state_rewards[3][0] = -1.0;
 gridWorld.state_rewards[3][1] = -1.0;
 gridWorld.state_rewards[3][2] = -1.0;
 gridWorld.is_terminal_state[0][0] = true;
 gridWorld.is_terminal_state[3][3] = true;
 gridWorld.initPolicy();
 gridWorld.PolicyIteration(0.5);

 gridWorld.is_wall[`y-axis`][`x-axis`] = true;