752f17f65ff851521b00f699e32f45e14945f069,tests/algorithms/test_td.py,,test_sarsa,#,102

Before Change

    pi, mdp, _ = initialize()
    alg = SARSA(pi, mdp.info, Parameter(.1))

    alg.Q.table = np.arange(np.prod(mdp.info.size)).reshape(mdp.info.size)

    alg._update(0, 1, 100, 1, 0)
    alg._update(1, 0, 10, 3, 1)
    alg._update(3, 1, 50, 3, 0)
    alg._update(2, 2, -100, 3, 1)

    test_q = np.array([[0, 11, 2, 3],
                       [4, 5, 6, 7],
                       [8, 9, -1, 11],
                       [12, 17, 14, 15]])

    assert np.array_equal(alg.Q.table, test_q)

After Change

    pi, mdp, _ = initialize()
    agent = SARSA(pi, mdp.info, Parameter(.1))

    core = Core(agent, mdp)

    // Train
    core.learn(n_steps=100, n_steps_per_fit=1, quiet=True)

    test_q = np.array([[4.31368701e-2, 3.68037689e-1, 4.14040445e-2, 1.64007642e-1],
                       [6.45491436e-1, 4.68559000, 8.07603735e-2, 1.67297938e-1],
                       [4.21445838e-2, 3.71538042e-3, 0., 3.439],

Instances