Iteration Results

Table of Contents

The file iteration_results.parquet (or iteration_results.csv) stores aggregate results for each iteration run by the simulator. The file is updated during the simulation, i.e., the results for an iteration are stored as soon as this iteration ends.

The format of this file is described in the table below. Many variables consist in four columns: the mean, standard-deviation, minimum and maximum of the variable over the population. A variable denoted as var_* in the table indicates that the results contain four variables: var_mean, var_std, var_min and var_max.

Column	Data type	Nullable	Description
`iteration_counter`	`Integer`	No	Iteration counter.
`surplus_*`	`Float`	No	Surplus, or expected utility, from the alternative-choice model (mean, std, min and max over all agents).
`trip_alt_count`	`Integer`	Yes	Number of agents traveling (agents who chose an alternative with at least 1 trip).
`alt_departure_time_*`	`Float`	Yes	Departure time of the agent from the origin of the first trip, in number of seconds since midnight (mean, std, min and max over all agents traveling).
`alt_arrival_time_*`	`Float`	Yes	Arrival time of the agent at the destination of the last trip, in number of seconds since midnight (mean, std, min and max over all agents traveling).
`alt_travel_time_*`	`Float`	Yes	Total travel time of the agent for all the trips, in seconds (mean, std, min and max over all agents traveling).
`alt_utility_*`	`Float`	Yes	Simulated utility of the agent for the selected alternative, departure time and route (mean, std, min and max over all agents traveling).
`alt_expected_utility_*`	`Float`	Yes	Expected utility, or surplus, for the selected alternative (mean, std, min and max over all agents traveling).
`alt_dep_time_shift_*`	`Float`	Yes	By how much the selected departure time of the agent shifted from the previous iteration to the current iteration, in seconds (mean, std, min and max over all agents traveling who chose the same alternative for the previous and current iteration).
`alt_dep_time_rmse`	`Float`	Yes	By how much the selected departure time of the agent shifted from the previous iteration to the current iteration, in seconds (root-mean-squared error over all agents traveling who chose the same alternative for the previous and current iteration).
`road_trip_count`	`Integer`	Yes	The number of road trips among the simulated trips.
`nb_agents_at_least_one_road_trip`	`Integer`	Yes	The number of agents with at least one road trip in their selected alternative.
`nb_agents_all_road_trips`	`Integer`	Yes	The number of agents with only road trips in their selected alternative.
`road_trip_count_by_agent_*`	`Float`	Yes	Number of road trips in the selected alternative of the agents (mean, std, min and max over all agents with at least one road trip).
`road_trip_departure_time_*`	`Float`	Yes	Departure time from the origin of the trip, in number of seconds after midnight (mean, std, min and max over all road trips).
`road_trip_arrival_time_*`	`Float`	Yes	Arrival time from the origin of the trip, in number of seconds after midnight (mean, std, min and max over all road trips).
`road_trip_road_time_*`	`Float`	Yes	Time spent on the road, excluding the time spent in bottleneck queues, in seconds (mean, std, min and max over all road trips).
`road_trip_in_bottleneck_time_*`	`Float`	Yes	Time spent waiting in a queue at the entry bottleneck of an edge, in seconds (mean, std, min and max over all road trips).
`road_trip_out_bottleneck_time_*`	`Float`	Yes	Time spent waiting in a queue at the exit bottleneck of an edge, in seconds (mean, std, min and max over all road trips).
`road_trip_travel_time_*`	`Float`	Yes	Travel time of the trip, in seconds (mean, std, min and max over all road trips).
`road_trip_route_free_flow_travel_time_*`	`Float`	Yes	Travel time of the selected route under free-flow conditions, in seconds (mean, std, min and max over all road trips).
`road_trip_global_free_flow_travel_time_*`	`Float`	Yes	Travel time of the fastest route under free-flow conditions, in seconds (mean, std, min and max over all road trips).
`road_trip_route_congestion_*`	`Float`	Yes	Share of extra time spent in congestion over the route free-flow travel time, in seconds (mean, std, min and max over all road trips).
`road_trip_global_congestion_*`	`Float`	Yes	Share of extra time spent in congestion over the global free-flow travel time, in seconds (mean, std, min and max over all road trips).
`road_trip_length_*`	`Float`	Yes	Length of the route selected, in meters (mean, std, min and max over all road trips).
`road_trip_edge_count_*`	`Float`	Yes	Number of edges on the selected route (mean, std, min and max over all road trips).
`road_trip_utility_*`	`Float`	Yes	Simulated utility of the trip (mean, std, min and max over all road trips).
`road_trip_exp_travel_time_*`	`Float`	Yes	Expected travel time of the trip, at the time of departure (mean, std, min and max over all road trips).
`road_trip_exp_travel_time_rel_diff_*`	`Float`	Yes	Relative absolute difference between the trip’s expected travel time and the trip’s actual travel time (mean, std, min and max over all road trips).
`road_trip_exp_travel_time_abs_diff_*`	`Float`	Yes	Absolute difference between the trip’s expected travel time and the trip’s actual travel time, in seconds (mean, std, min and max over all road trips).
`road_trip_exp_travel_time_diff_rmse`	`Float`	Yes	Absolute difference between the trip’s expected travel time and the trip’s actual travel time, in seconds (root-mean-squared error over all road trips).
`road_trip_length_diff_*`	`Float`	Yes	Length of the selected route that was not selected during the previous iteration (mean, std, min and max over all road trips for agents who chose the same alternative for the previous and current iteration).
`virtual_trip_count`	`Integer`	Yes	The number of virtual trips among the simulated trips.
`nb_agents_at_least_one_virtual_trip`	`Integer`	Yes	The number of agents with at least one virtual trip in their selected alternative.
`nb_agents_all_virtual_trips`	`Integer`	Yes	The number of agents with only virtual trips in their selected alternative.
`virtual_trip_count_by_agent_*`	`Float`	Yes	Number of virtual trips in the selected alternative of the agents (mean, std, min and max over all agents with at least one virtual trip).
`virtual_trip_departure_time_*`	`Float`	Yes	Departure time from the origin of the trip, in number of seconds after midnight (mean, std, min and max over all virtual trips).
`virtual_trip_arrival_time_*`	`Float`	Yes	Arrival time from the origin of the trip, in number of seconds after midnight (mean, std, min and max over all virtual trips).
`virtual_trip_travel_time_*`	`Float`	Yes	Travel time of the trip, in seconds (mean, std, min and max over all virtual trips).
`virtual_trip_global_free_flow_travel_time_*`	`Float`	Yes	Minimum travel time possible for the trip, in seconds (mean, std, min and max over all road trips). Only relevant for time-dependent virtual trips.
`virtual_trip_global_congestion_*`	`Float`	Yes	Share of extra time spent in congestion over the global free-flow travel time, in seconds (mean, std, min and max over all road trips). Only relevant for time-dependent virtual trips.
`virtual_trip_utility_*`	`Float`	Yes	Simulated utility of the trip (mean, std, min and max over all virtual trips).
`no_trip_alt_count`	`Integer`	No	Number of agents not traveling (agents who chose an alternative with no trip).
`sim_road_network_cond_rmse`	`Integer`	Yes	Root-mean-squared error between the simulated edge-level travel-time function for the current iteration and the expected edge-level travel-time function for the previous iteration. The mean is taken over all edges and vehicle types.
`exp_road_network_cond_rmse`	`Integer`	Yes	Root-mean-squared error between the expected edge-level travel-time function for the current iteration and the expected edge-level travel-time function for the previous iteration. The mean is taken over all edges and vehicle types.

METROPOLIS2 The Book

Iteration Results