The data structure used to store information about a single rollout. More...
#include <utils.h>
Public Attributes | |
Eigen::MatrixXd | control_costs |
A matrix [num_dimensions][num_time_steps] of the control cost for each parameter at every timestep. | |
std::vector< double > | full_costs |
A vector [num_dimensions] of the full coss, state_cost + control_cost for each joint over the entire trajectory full_costs_[d] = state_cost.sum() + control_cost[d].sum() | |
std::vector< double > | full_probabilities |
A vector [num_dimensions] of the probabilities for the full trajectory. | |
double | importance_weight |
importance sampling weight | |
Eigen::MatrixXd | noise |
A matrix [num_dimensions][num_time_steps] of random noise applied to the parameters. | |
Eigen::MatrixXd | parameters_noise |
A matrix [num_dimensions][num_time_steps] of the sum of parameters + noise. | |
Eigen::MatrixXd | probabilities |
A matrix [num_dimensions][num_time_steps] of the probability for each parameter at every timestep. | |
Eigen::VectorXd | state_costs |
A vector [num_time_steps] of the cost at each timestep. | |
double | total_cost |
combined state + control cost over the entire trajectory for all joints | |
Eigen::MatrixXd | total_costs |
A matrix [num_dimensions][num_time_steps] of the total cost, where total_cost[d] = state_costs_ + control_costs_[d]. |
The data structure used to store information about a single rollout.