Constant traffic growth requires higher levels of safety and traffic management. Passenger transport control covers many business processes that can be split into several groups: planning, supervision, control, analysis, paperflow management, workforce management. The paper deals with modeling and automation analysis of the Moscow metro traffic, which covers 56% of passenger throughput in one of the world's largest cities.
Constant traffic growth demands for higher levels of safety and traffic management
quality. The unified automated information control system of planning and
management of transportation process (AICS TP) copes with the challenge. The system forms
unified information space and uses modern processing know-how. The paper contains
authors' experience in modeling and automation of Moscow metro business processes.
Moscow metro covers 56% of passenger throughput in one of the world’s largest
cities. Traffic management is carried out by the Dispatch Service (DS). In the eighties
of the last century there were attempts to create automated planning tools for traffic
management purposes [
Until 1998 the DS implemented both in-house (developed by employees of the information-computing center of the Moscow underground) and outsourced automated systems. These were systems for the remote management of objects on the metro lines and means for information visualization.
In 1998 Moscow government supported the start of the metro traffic complex automation. The initial target was to automate DS operating staff training process. Training consists of games that test the ability to manage traffic in case of failure and to mitigate any failure effect according to existing job descriptions.
Until that time training games were carried out as follows. The coach and the learner took part in a training game. Everyone had the same sheet of planned schedule. The coach specified situations and violations, simulated messages from workers lines, accepted commands from the learner, wrote all information incoming and outgoing from him on the sheet of planned schedule and drew current position of trains in real time. The learner took simulated messages and responded to them, giving a voice command, wrote accepted messages and commands specifying the time on his sheet of planned schedule and also drew current position of trains. According to the results of the game, the coach estimated the actions of the learner due to its completeness and timeliness of management.
In 2000 the task was set to automate train movement planning. Until then trains were scheduled manually. At the same time as a result of gained data and analysis, the goal was set to automate passenger trains control process. These related tasks can be efficiently solved within the single AICS TP. 3
First of all we defined a set of system users and their tasks [
We were looking to establish a set of processes for the unit that directly exercises control over transport movement. Each business process corresponds to a set of separate functions. Normally these functions are executed by the unit as a whole. Table 1contains this information.
Fig. 1 shows upper level hierarchy of the planning and control related business processes. Process identification allowed us to single out AICS TP main components, which will be responsible for process automation:
Detailed description drafting the schedule of passenger vehicles (hereinafter, the
schedule); planning adjusted to the schedule technical maintenance
works of the transport company,; planning technical means to reinforce schedule (scheduled
and operative, in case of deviations from planned). reinforcing the schedule; providing traffic safety; assuring movement control when there are deviations from
the schedule; adopting operational measures to restore the traffic on
schedule in the shortest possible time; adopting operational measures to ensure passenger trans
portation in case of considerable passenger traffic increase; using to the utmost existing infrastructure to provide for the
specified amount of traffic.
Control over the work of other departments analyzing the schedule; estimating actions of employees, connected with movement
of vehicles; calculating technical and economic indices of the opera
tional work of the entity. providing information on the schedule; registering schedule deviations; providing information on schedule deviations; exchanging operational information with other units’ em
ployees. providing human resources; improving professional skills and personnel training; providing labor protection items; carrying out measures on civil defense plan.
Other functions determined by internal documentation of the company and instructions of the company management. the automated system of the planned schedule construction; simulator of a train dispatcher; the automated system of operative dispatch management; automated system of the objects on the subway lines remote management; means to visualize objects status and current trains position information.
One of the main components of the AICS software is the multifunctional model of
subway line developed with the participation of the authors [
The model includes the following main components: the line model; train traffic control model; trains model; energy supply systems model; turnouts and signals control system model; traffic interval regulation model; passenger traffic model; tasks management model; results evaluation model; simulation results visualization model.
Separate technical facilities models include: algorithm of the proper object functioning; description of failures which may arise during the object operation; description of the failures’ impact on the state of the object.
This provides the modeling of standard and non-standard technological situations. Depending on the purposes for which multifunctional model is used, the level of detail and construction type of individual model can be customized. 4.2
Next we constructed solving tasks scenarios with detailed automated business processes. As an example of the detailed top level business process we will look at the technical means control process. This process is employed primarily to maximize the use of existing infrastructure to support the specified traffic level. This is one of the main activities for managing the movement of passenger vehicles. Fig. 2 presents the model of such business process constructed with regard to the underground Dispatch Service. DS employees as well as employees of other services directly managing technical means (persons on duty of centralization posts, stations, depot, train drivers) who report operationally to train dispatcher, are users of this process.
Let us consider how the actions performed by the Dispatcher-centralizer in the process of operation of turnouts and signals with automated working places (AWP) affect the operation of technical means. As a rule, dispatcher-centralizer implements process control with the set of relays, turnouts or signals. The management process of control of individual relays, turnouts, signals is implemented within the control process of set of relays, turnouts or signals, or, in case of deviations from the schedule of movement, functional disorders of technical means, diagnosing their functioning.
The basic process of control of set of relays, turnouts or signals is the route of
route-relay interlocking (hereinafter, the route) management process. The route is
called the pre-prepared way with turnouts, translated and secured in the correct
position for the intended train movement or shunting composition [
The route has multiple states, in each of which some actions are performed. The
transition from one state to another is carried out under certain conditions.
Formalization of description of the functioning of such systems is done conveniently using Petri
nets [
Then each state corresponds to the place of Petri network. Each transition of a system from one state to another corresponds to the Petri net transition, which is the fulfillment of certain conditions. Also the advantage of this formalization is a marking that uniquely identifies the network status. The vertices of the graph are divided into two sets and connected so that each arc is directed away from the element of one set to the element of another set. This allows us to clearly describe the features of functioning of routes. Fig. 3 shows a network in which places correspond to route states: p0m – the route is inactive; p m
1 – the route is specified; p m
2 –route is assembled – turnouts were translated to the correct position; p m
3 – the route is preliminarily locked – section before the route is free and indication of the route is turned on; p m
4 – the route is finally locked – section before the route is busy; p m
5 – artificial route cutting.
Fig. 4 presents the formalization of the same process, carried out using the EPC notation used to describe business processes.
The choice of the particular description form depends on the purpose for which this
description is used. If the primary object is technical means and, for example, the task
of modelling of its functioning is solved, then the first form is more convenient. Using
Petri nets approach for modelling the process of route control presented in this paper,
has shown its effectiveness when functioning of various modifications routes and
their associations (automatic route mode) needs to be taken into account. One of the
goals for constructing train dispatcher simulator is to train infrastructure control
process. Therefore these questions were specifically addressed when train dispatcher
simulator was created. At the same time Petri net proved its effectiveness in the model
of train traffic control [
If processes are being depicted in employees’ job description or for training purposes (shaping learning scenarios), the second form is more convenient. 4.2
AICS software automates scheduling and subway trains control. Regularity of the
intervals of motion is one of the key principles to be implemented for the purpose of
these tasks. It's essential from the point of view of quality of passenger services. If
number of trains on the line is unchanged then all traffic intervals should be equal.
When the number of trains on the line is changed in accordance with passenger traffic
fluctuations, traffic intervals are not equal. Irregular intervals cause inconvenience for
passengers. The degree of discomfort is estimated duration of the transition process
alignment intervals. The decline of such discomfort can be achieved by a uniform
entry/removal of trains. It provides the minimum duration of the transition process for
subsequent alignment intervals [
Let us consider results of AICS TP components and others automation tools implementation, developed by the Department of "Management and information Security" of the Moscow State University of Railways Engineering (MIIT) and for many years used in the Moscow metro.
The simulator train dispatcher [
The planned schedule construction automated system is used for automation of
planning traffic processes, providing information about the timetable, calculation of
technical and economic indicators of the operational work of the company [
The automated system of operative dispatch management can be used for
automation of management processes movement in case of deviation from the planned
schedule, restoration movement on planned schedule, calculation of technical and
economic indicators of the operational work of the company, performance analysis of
the planned schedule and information on schedule deviations [
Multifunctional model of subway line can be also applied for other business
processes (technological processes) modeling tools in the Moscow metro, such as:
optimal energy consumption calculation automated system is used to automate the
process of technical means planning [
Presented in the case-study paper the automated systems are applied to the conditions of all lines of the Moscow metro, use a common database and include subsystems for data conversion, which allow to organize automatic information exchange. All applied systems are constantly updated to reflect the new demands of the Moscow metro caused by its development and growing requirements for quality and volume of services.
Comprehensive approach to the modelling of business-processes allows to automate many processes of movement control with minimal labor and financial costs while making maximum use of existing data and knowledge databases. 6
Developed in the beginning of XXI century software and purchased at the same time hardware are outdated morally and physically and require reengineering based on accumulated experience.
High rates of development of the Moscow metro urge to automate the process of creating and editing model elements, determine the relationships between objects, develop graphical forms display, write games scenarios. This enables to speculate on the need to create automated model designing system or specialized CASE-tools.
Within automated planning and trains movement management, the choice of the control decisions sequence is performed using recursive procedures. This approach helps to obtain a wide set of variants of planned and operative schedules. The formation of this set and selection of a rational solution can be speeded up with the use of parallel programming methods. These challenges and radical changes in system software require the development of new application software in modern programming environments.