Simulation Mechanism: Summary

In this page, we briefly explain the models and logics behind UXsim. If you want to know more details, please refer to arXiv preprint on UXsim, the code, the original articles, or Japanese textbook on traffic simulation.

Models

UXsim computes network traffic flow dynamics by combining following models:

• The mesoscopic version of Newell’s simplified car-following model (aka. X-model) for link and vehicle dynamics.

• The mesoscopic version of the incremental node model for inter-link transfer at nodes.

• Dynamic user optimum (aka. reactive assignment) with stochasticity and delay for route choice of travelers.

Key Inputs

UXsim’s key inputs (i.e., simulation scenario parameters) are as follows:

• Reaction time of vehicles (reaction_time argument of World). This value determines the simulation time step width. This is a global parameter.

• Platoon size for mesoscopic simulation (deltan of World). This value determines the simulation time step width. This is a global parameter.

• Route choice model parameters: shortest path update interval (duo_update_time of World) and weight value (duo_update_weight of World).

• Lists of nodes and links. They define the network structure.

• Parameters of each link. For example, length, free flow speed (free_flow_speed), number of lanes (number_of_lanes), jam density (jam_density), merging priority parameter (merge_priority), and traffic signal setting (signal_group).

• Parameters of each node. For example, position (x and y) and traffic signal setting (signal).

• Demand. For example, origin, destination, and departure time of each vehicle are specified.

Program Structure

Class diagram of UXsim: the relationship between classes and their key properties.

Computation Procedure

Activity diagram of UXsim: the overall computation procedure of UXsim.

Activity diagram of a specific vehicle in UXsim: the computation procedure for each vehicle.

Traffic Flow Model Parameters

The following parameters are used in UXsim, defined through Link, World, and Node (*: input, **: alternative input). If you are not familiar with the traffic flow theory, it is recommended that you adjust only free_flow_speed and number_of_lanes for the traffic flow model parameters, leaving the other parameters at their default values.

• * free_flow_speed (m/s)

• * jam_density (veh/m/LINK)

• ** jam_density_per_lane (veh/m/lane)

• * lanes, number_of_lane (lane)

• tau: y-intercept of link FD (s/veh*LINK)

• REACTION_TIME (s/veh*lane)

• w (m/s)

• capacity (veh/s/LINK)

• capacity_per_lane (veh/s/lane)

• delta: minimum spacing (m/veh*LINK)

• delta_per_lane: minimum spacing in lane (m/veh*lane)

• q_star: capacity (veh/s/LINK)

• k_star: critical density (veh/s/LINK)

• * capacity_in, capacity_out: bottleneck capacity at beginning/end of link (veh/s/LINK)

• * Node.flow_capacity: node flow capacity (veh/s/LINK-LIKE)