hiv.py 文件源码

def perform_action(self, action, perturb_params=False, p_lambda1=0, p_lambda2=0, p_k1=0, \
        p_k2=0, p_f=0, p_m1=0, p_m2=0, p_lambdaE=0, p_bE=0, p_Kb=0, p_d_E=0, p_Kd=0, **kw):
        """Perform the specifed action and upate the environment.

        Arguments:
        action -- action to be taken

        Keyword Arguments:
        perturb_params -- boolean indicating whether to perturb dynamics (default: False)
        p_lambda1 -- hidden parameter (default: 0)
        p_lambda2 -- hidden parameter (default: 0)
        p_k1 -- hidden parameter (default: 0)
        p_k2 -- hidden parameter (default: 0)
        p_f -- hidden parameter (default: 0)
        p_m1 -- hidden parameter (default: 0)
        p_m2 -- hidden parameter (default: 0)
        p_lambdaE -- hidden parameter (default: 0)
        p_bE -- hidden parameter (default: 0)
        p_Kb -- hidden parameter (default: 0)
        p_d_E -- hidden parameter (default: 0)
        p_Kd -- hidden parameter (default: 0)
        """
        self.t += 1
        self.action = action
        eps1, eps2 = self.eps_values_for_actions[action]
        r = ode(self.model_derivatives).set_integrator('vode',nsteps=10000,method='bdf')
        t0 = 0
        deriv_args = (eps1, eps2, perturb_params, p_lambda1, p_lambda2, p_k1, p_k2, p_f, p_m1, p_m2, p_lambdaE, p_bE, p_Kb, p_d_E, p_Kd)
        r.set_initial_value(self.state, t0).set_f_params(deriv_args)
        self.state = r.integrate(self.dt)
        reward = self.calc_reward(action=action)
        return reward, self.observe()