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

Usage

calculateBellmanWithIterativeSimulations(
  area,
  pumping,
  pump_eff = 1,
  opts,
  nb_control = 10,
  nb_itr = 3,
  mcyears,
  penalty_low,
  penalty_high,
  path_solver,
  states_step_ratio = 1/50,
  cvar_value = 1,
  penalty_final_level = NULL,
  df_previous_cut = NULL
)

Arguments

area

Area with the reservoir

pumping

Binary, TRUE if pumping is possible

pump_eff

Pumping efficiency (1 if no pumping)

opts

List of simulation parameters returned by the function antaresRead::setSimulationPath

nb_control

Number of controls used in the interpolation of the reward function

nb_itr

Max number of iterations

mcyears

Vector of years used to evaluate rewards

penalty_low

Penalty for violating the bottom rule curve, comparable to the unsupplied energy cost

penalty_high

Penalty for violating the top rule curve, comparable to the spilled energy cost

path_solver

Character containing the Antares Solver path, argument passed to runSimulation.

states_step_ratio

Discretization ratio to generate steps levels between the reservoir capacity and zero

cvar_value

from 0 to 1. the probability used in cvar method

penalty_final_level

Penalties (for both bottom and top rule curves) to constrain final level

df_previous_cut

Data frame containing previous estimations of cuts

Value

List containing aggregated water values and the data table with all years for the last iteration