Semillas aleatorias en modelos de transporte
Esta nota técnica explicará las diferentes fuentes de estocasticidad dentro de los micro y meso experimentos en Aimsun Next y cómo puedes controlarlas.
How to Use Hybrid Macro-Meso Simulation in Aimsun Next 20
Technical Note #50
August 2020
By Tessa Hayman
Tessa Hayman explains hybrid macro-meso simulation in Aimsun Next 20 and how it can unite the best of macro and meso to enable regional-scale modeling.
Traditionally, the strategic and operational impacts of a transportation scheme or future plan are analyzed by building two separate models: one dealing with planning, and the other with operations. The hybrid macro-meso adopts the innovative approach of fusing both in a single model.
How does it work?
Unlike traditional macroscopic modeling, the hybrid macro-meso simulator is a dynamic model that simulates individual vehicles. The paths of each vehicle are determined by a dynamic assignment. The main difference between the macroscopic and mesoscopic levels is in the way that a vehicle’s travel time is calculated: network loading.
The network loading and path assignment within Aimsun Next are treated separately, which allows for hybrid modeling.
The network loading gives the travel time and number of vehicles assigned on a section to the route calculation every route choice interval.
The cost of a path is: Macro cost + Meso cost
Simulating individual vehicles
We can split the simulation of a vehicle into 3 main steps:
- Generating a vehicle
- Calculating a route
- Simulating the route
The vehicle generation is the same as in mesoscopic simulation: the modeler chooses an arrivals model.
The route calculation is the same as in mesoscopic simulation. The modeler chooses either stochastic route choice or dynamic user equilibrium and the cost of each path is the macro cost + the meso cost.
The route is then simulated as follows:
1. If the origin is in the macro area, calculate the delay functions to give a macro travel time.
2. Move the vehicle to the turn that enters the meso area (either instantaneously or with a delay from the functions).
3. Add one vehicle to the volume of the sections traversed.
4. Can the vehicle safely enter the meso section?
Yes – simulate as a mesoscopic vehicle;
Add to the virtual queue.
If the destination is in the macro area
- At the vehicle’s exit time from the meso area, calculate travel time using the delay functions;
- Add one vehicle to the volume of the sections traversed;
- Move vehicle to destination OR to next mesoscopic area.
How to create a hybrid macro-meso experiment
Within a dynamic scenario, you can create a hybrid macro-meso experiment in a similar way to creating a hybrid meso-micro experiment.
- Draw a polygon around each area that you would like to simulate at a mesoscopic level
- Convert the polygon to a simulation area
- Create a new hybrid macro-meso experiment within an existing dynamic scenario
- In the Hybrid tab of the new experiment, add your polygon(s) to specify the mesoscopic areas
Draw a polygon, right click and click convert to simulation area
The polygon will then go green to show it is a simulation area
Cost alignment
As the path cost is the sum of the macro cost and the meso cost, it is essential that the units of the macro and meso cost functions are aligned. The default macroscopic cost is expressed in minutes, whereas the default mesoscopic cost is expressed in seconds. To align the cost, you can specify the conversion in the hybrid macro/meso experiment, using the VDF/TPF/JDF Unit Conversion Factor.
It is a good idea to check that the functions that you have used are the same format as the generalized cost function. For example, you may need to add in toll costs, value of time and vehicle operating costs with the same coefficients.
If you want to get skim matrices with any function components, you’ll also need to ensure that they are defined both in your macroscopic functions and in your mesoscopic functions.
Travel time in the macro area
The travel time in the macro area specifies how long it takes for a vehicle generated in the macroscopic area to reach the boundary of a mesoscopic area. The default is Total VDF/TPF/JDF cost, which interprets the overall output of those functions as a pure travel time; if you have added other terms and computed a generalized cost, you should define and choose a function component that provides the travel time only.
If you developed your demand by using a cordon around the mesoscopic area, you should choose instantaneous time for the first and last segment. This is because the time of each demand slice represents the time at which incoming trips cross the boundary of the mesoscopic area rather than the time when they departed from their origin.
Más notas técnicas
Controlador C/ACC en Aimsun Next
Octubre de 2021: En este artículo, Martin Hartmann explica y demuestra la microsimulación de Aimsun Next de vehículos equipados con control de crucero adaptativo cooperativo (CACC).
