Calculate Bellman values throughout iterations of Antares simulation and DP Each simulation leads to a new reward estimation, which leads to new water values, which leads to the off-line calculation in R of an optimal trajectory, which leads to new controls to be evaluated which leads to a new simulation
Calculate Bellman values throughout iterations of Antares simulation and DP Each simulation leads to a new reward estimation, which leads to new water values, which leads to the off-line calculation in R of an optimal trajectory, which leads to new controls to be evaluated which leads to a new simulation
Compute reward function with the 5 simulations method. Called for each area of MultiStock_H2_Investment_reward_compute_once. If they are several areas, the trajectories of the other storage are fixed on their optimal trend.
Compute reward function for node with plaia implementation. The reward function of each week and each year is evaluated on nb_simulations controls. Called for each area of getBellmanValuesSequentialMultiStockWithPlaia(). If they are several areas, the trajectories of the other storage are fixed on their optimal_trend.
Get objective values of the optimization problem of each week and each scenario for a given simulation simu, mainly used in get_Reward() to build reward functions.
Computes the total reward for a list of candidate following the reward function, with parallel processing. The storage volume is constant. These function is called inside parLapply().