ModeLab-Web Tool for the Modeling of Bus Rapid Transit Systems

Bus rapid transit (BRT) systems have in recent years become a viable and effective option for solving the mobility problems in different cities around the world. For these systems to fulfill their mission effectively and efficiently, they need to respond adequately to the different situations that appear periodically or arbitrarily in the users’ routines, modification of system resources (e.g., buses, drivers, lanes, or roads), among others. Modeling and simulation in these and many other complex systems are key tools to support decision-making since, in general, they are an inexpensive option that allows to quickly evaluate the effect of different 1 Universidad del Cauca (Popayán-Cauca, Colombia). crcampo@unicauca.edu.co. ORCID: 0000-0002-2395462X 2 Universidad del Cauca (Popayán-Cauca, Colombia). ORCID: 0000-0001-5405-0247 3 Ph. D. Universidad del Cauca (Popayán-Cauca, Colombia). ccobos@unicauca.edu.co. ORCID: 0000-00026263-1911 ModeLab Web Tool for the Modeling of Bus Rapid Transit Systems Revista Facultad de Ingeniería (Rev. Fac. Ing.) Vol. 30 (56), e13294. April-June 2021. Tunja-Boyacá, Colombia. L-ISSN: 0121-1129, e-ISSN: 2357-5328. DOI: https://doi.org/10.19053/01211129.v30.n56.2021.13294 changes on the system and to define the best solution in the shortest time for a specific problem or situation. This paper introduces ModeLab, a web-based tool for modeling BRT systems that facilitates the design of models by using an iconic language closer to the modeler, a language based on the real-world objects found in this type of system and that allows us to define simpler and more compact models, which are easier to visualize, understand, and configure. To evaluate the models that ModeLab can define, a model of medium complexity was developed and compared with the model obtained by ARENA®; the results show a significant reduction in the complexity of the models, while, at the same time, there are identical results when simulating the models with SIMAN (a common simulation software for both tools).


I. INTRODUCTION
The vast majority of cities in the world are growing in the number of inhabitants at an unprecedented pace [1]. This requires, among other things, urban transport solutions that can be quickly implemented and allow efficiently transporting large volumes of passengers. Massive passenger transport systems (MPTS) or Bus Rapid Transit (BRT) Systems have proven to be a viable solution to the growing demand for public transport and the current problems of mobility [2][3][4] [5].
The implementation of the MPTS involves complex design and programming problems, among which are the design of networks and transport routes, frequency programming, and schedules scheduling, which are part of a global problem called "Transit Network Design and Scheduling Problem" (TNDSP) [6] [7]. Solving these problems generally involves the use of modeling and simulation tools that allow evaluating the impact of changes in the system to make more appropriate decisions to improve the service.
In the market, there are different modeling and simulation tools, including ARENA® Simulation Software [8][9][10] [11]. A commercial tool to model the operation of any system based on the concept of discrete events and supported in SIMAN [12] [13] to perform the simulation. ARENA® acts as an automatic SIMAN code generation tool. However, its use brings two drawbacks. On the one hand, the high licensing costs, which can directly affect the viability and start-up of a project, and on the other, the fact that it requires very specialized knowledge to make the models.
Modeling a full MPTS in ARENA® can become a complex task because, being a general domain tool, it uses abstract modeling elements that are difficult to relate to real-life objects, and even more so with specific elements of an MPTS (Buses, Stations, Passengers, Intersections, among others). This was the impulse behind the creation of a specific tool (ModeLab), whose interface includes an iconic language (icons and objects) directly related to the form and behavior of the objects in the actual MPTS, which saves time as it facilitates to designers and experts the construction of the models. In addition, this software is free to access The elements that make up the Megabus MPTS include Bus depots, points in the system where the departure and arrival of articulated buses are controlled and the routes that make the journey through the system are managed; Stations (bus stations), facilities within the system where passengers enter or leave the system or where they board or disembark from buses (double stations allow passengers to transfer from a bus that goes in one direction to another, and the simple ones that only have mobility in one direction); Routes, which determine the entire journey of the buses; in the case of the trunk routes, these determine only the journey of the buses between the stations; Buses, the vehicles with the ability to transport a maximum of passengers; in the case of articulated buses, these only go by exclusive lanes making stops between stations; Intersections, points of road infrastructure and transport where the road is divided into two or more paths; and Journey time, which corresponds to the time it takes a bus to go from one station to another. Fig. 1 shows an abstract and complete model of Megabus.

III. MODEL OF MEGABUS IN ARENA®
ARENA® stands out for its versatility to model several domains, a relatively low learning curve and a high level of interoperability with other applications using the SIMAN compiler and simulator. Different ARENA® modules were used for the Megabus modeling, namely: • Data modules, which allow defining storage structures necessary for the implementation of the simulation. They include entities (people, objects, etc.), attributes (a common feature of entities), and variables (global to the system and accessible by all modules).    The Bus Station was organized in a sub-model that does not differentiate whether it is single or double. This sub-model has the following modules: Entry (Decide), which validates whether the bus that reaches the station enters or continues its route to another station; Wait (Hold), which is responsible for queueing the buses that are coming to a station in case the station is busy; Register (Record), which registers the entry of a bus to a station; Load (Process), which allocates the bus a fixed time that simulates the loading and unloading of passengers; and Departure (Assign), which records the bus departure and assigns the new capacity to the bus after having collected and dropped off passengers, and it is informed that the station is free to be occupied by other buses. Fig. 3 (b) shows this sub-model. More Options (Decide), which verifies whether the passenger has more options to take their route -if the answer is positive, the passenger passes the "Next Option" module, otherwise, they go to the "Wait for another bus" module-; Next Option (Assign) is responsible for assigning the following travel options to the passenger to allow them to evaluate all their travel possibilities against another bus; Wait for Another Bus (Hold), which queues a passenger if they cannot board the bus that is at the station (there is no room or it is the wrong route) until the bus leaves the station; Arrival (Decide), which consults whether the passenger arriving has the present station as their destination or a stop -if the answer is positive, they go to the "Disembarkation" module, otherwise, they continue their journey-; Arrive The Double passenger station has two entrances and two exits. The additional entrance and exit models the transfer of a passenger from one station to another.
The modules added to single passenger stations are two: Re-entry Point (Decide) queries whether the next passenger stop is the adjoining station -if so, the passenger passes to the "Travel to E#" module; otherwise, they wait for the bus at the station-; and Travel to E# (Assign), which saves in Passenger.TotalJourney the change of station made by the passenger. Fig. 4 shows this sub-model.

IV. MODEL OF MEGABUS IN MODELAB
ModeLab is a specialized Web tool for the construction of MPTS simulation models; it seeks to reduce the complexity in the preparation of this type of models when using the existing tools and reduce the high licensing costs of such tools.  Tables like them were developed, featuring copy, paste, undo, and redo functionalities, which also allow validation, classification, and data grouping. These tables support the inclusion of formulas in columns, rows, or cells. These types of tables were also used for configuration of vector and data modules (routes, probability of destinations, short routes) allowing copying data from other formats (txt, xls, doc) directly on tables, which constitutes a very important aspect for the tool since, for example, the short route matrix is generated from an external algorithm or manually in a text file with all possible short routes.
Depending on the number of stations, this matrix can become very extensive (NxN), and passing this data to the one-in-one tool can become an exhausting task.  Allows to zoom out or approach the canvas by using the mouse scroll wheel or the toolbar.

Drag & Drop
The action of being able to drag the items found in the left bar to the modeling area (canvas).

Save Information
Through emerging modal windows, it is possible to store data on the items and maintain synchronized information with the tables.
Each of the elements of the model has attributes that must be configured by the user. By way of example,    To be able to compare the level of reduction of complexity of the models and ensure that simulation results with SIMAN were equal, a model was developed of a simple system with 6 stations (3 double stations) and 1 bus depot. Fig. 7 shows the model in the ModeLab tool, linearly represented in a very short space and easy to understand. Meanwhile, Fig. 8  can be seen at first glance that the second model is more complex and difficult to understand. The results of the simulator for the two models were identical.

V. DISCUSSION AND CONCLUSIONS
With the development of this work, the modeling and implementation of a Web tool (ModeLab) were achieved, which allows the construction of MPTS models more naturally and intuitively compared to other modeling and simulation tools on the market. Specifically, it is easier to make MPTS models this way than with ARENA®. Moreover, ModeLab also makes it possible to export the model to SIMAN scripts, which can be executed and thus obtain the results of their simulation.
As future work, the research group expects to refine the Megabus model since the passenger arrival parameters, the destination of the parameters, and others must be calibrated or adjusted to various periods of the day. In addition, it is expected to include in ModeLab some functionalities that allow the management and storage of different simulation models in a centralized repository and to include the simulation of the model in the same application.

AUTHOR'S CONTRIBUTION
Carlos-Robinson Campo: research, data collection, methodology, formal analysis, original draft writing, writing revision and editing.
Juan-Pablo Salazar: research, data collection, methodology, formal analysis, original draft writing, writing revision and editing.

ACKNOWLEDGMENTS
The work presented in this paper was partially supported by the Information Technology Research and Development Group (GTI) of the University of Cauca.
We are grateful to Colin McLachlan for assisting with the translation of the paper.