Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Contents lists available at ScienceDirect

Transportation Research Interdisciplinary Perspectives

journal homepage: www.sciencedirect.com/journal/transportation-
research-interdisciplinary-perspectives

Timetable rationalization & Operational improvements by human

intervention in an urban rail transit system: An exploratory study
Manish Kumar Sharma a, *, Bhavesh Kumar Chauhan b
a
Sr. Executive - Operations, National Capital Region Transport Corporation, Delhi, India
b
Professor, Shri Ramswaroop College of Engineering & Management, Lucknow, Uttar Pradesh, India

A R T I C L E I N F O A B S T R A C T

Keywords: The efficient scheduling of train running is fundamental to operational planning of train movement in the
Demand passenger-based transportation industry. The timetable determines the availability of transport services across
Supply time & space. It also indicates the need of supporting infrastructures like rolling stock, train control system,
Timetable
traction system, and manpower. The rationalization of the timetable may profoundly affect the efficiency of the
Train operations
Headway
operator by optimizing the need for supporting infrastructure. Timetable rationalization in an urban rail transit
system is considered with the two aims, first is human intervention to meet up the rising passenger demand in
form of ridership. Second is lowering the operational expenditure in terms of energy consumption during train
operation. A data-based exploratory case study of Delhi Metro line-2 is used to demonstrate the performance of
the proposed approach. This paper presents an overview of the initiatives adopted in Delhi Metro Rail Corpo­
ration to ameliorate the train schedule and timetable by adopting innovative ways like matching supply with
demand, real-time monitoring of demand for necessary intervention in the timetable of the day, coasting or
energy-saving mode of operations in Automatic Train Operation, dwell time optimization, and interlacing of
high-capacity trains during peak hours.

Introduction viable & sustainable including induction & dispatch of trains from/to
depot without compromising the financial sustainability of the urban
Timetabling is a passenger-centric act of balancing demand with the transit operator.
supply of transportation services to ensure adequate capacity to clear the In rail-based urban transit of metro cities, for service quality im­
passenger rush during a particular moment with optimal utilization of provements different operational strategies like adjustment in metro
coaches and assets. The demand for transportation is characterized by train speed (Feng et al., 2011), optimizing coaches composition (Feng
the direction of travel, time of travel, origin & destination points, et al., 2013)and introduction of big-data-based methodology encour­
interchange stations. The supply of transportation services is measured aging decision making of both operators & travelers (Taboada and Han,
in terms of the availability of accommodation in the vehicle from one 2020)are advised to go through. Among them, designing the timetable
place to another during any hour of the day (Ceder et al., 2001). If the as per commuter flow is accepted as the most effective (Sun et al., 2014).
supply of these services is less than the demand, it would cause incon­ The timetable designing with a mathematical model is most promising
venience to the commuters whereas the reverse would result in subop­ among researchers (Carey and Crawford, 2007; Castillo et al., 2009;
timal utilization of assets and would lead to high costs to the urban Castillo et al., 2011; Castillo et al., 2015). Many researchers introduced
transit operator (Robenek et al., 2016). In technical terms, a timetable is optimization problems in timetable designing for only a line (Cury et al.,
a kind of planning & listing the time at which arrivals & departures of 1980; Furth and Wilson, 1981; Wang et al., 2013; Barrena et al., 2014)
trains at a station, are expected to take place ensuring adequate coaches and a complete railway network (Palma and Lindsey, 2001; Caprara
availability as per passenger convenience and keeping the operations et al., 2002; Liebchen, 2008; Wong et al., 2008; Aksu and Akyol, 2014)

Abbreviations: AFC, Automatic Fare Collection; ATO, Automatic Train Operation; DMRC, Delhi Metro Rail Corporation; DRT, Demand Responsive Transport; HCC,
Huda City Centre; JGPI, Jahangirpuri; PHPDT, Peak Hour Peak Direction Traffic; PIDS & PAS, Passenger Information Display System & Passenger Announcement
System; QM, Qutubminar; SPBI, Samaypur-Badli; VW, Vishwavidyalaya.
* Corresponding author.
E-mail addresses: manish.dtutech@gmail.com (M. Kumar Sharma), blessu35@gmail.com (B. Kumar Chauhan).

https://doi.org/10.1016/j.trip.2021.100526
Received 9 June 2021; Received in revised form 13 December 2021; Accepted 18 December 2021
Available online 3 January 2022
2590-1982/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

but none of them considers the factor of the capacity of metro trains & This kind of study has never been done before in any metro system of
stations. Later on, in search of optimized timetable metro train capacity, India. It is first of its kind original work. The findings are described in
ridership, train frequency, variable headway, time-dependent dwell results as cost optimization in form of energy saving cumulatively
time etc. constraints are taken into consideration (Ceder, 1984; Ceder, meeting train supply with passenger demand.
2001; Ceder, 2007; Koutsopoulos et al., 1985; Boroun et al., 2020). Here
commuter’s platform waiting time from station entry to board the train The demand for urban transport services
& first come first serve or Queue pattern is also important in timetable
optimization models (Albrecht, 2009; Chen, 2014; Hassannayebi et al., The need for transportation is generated due to differential place­
2014; Niu and Zhou, 2013; Wu et al., 2015) with a reduction in net ment of demand and supply of goods & services in time & space. People
energy consumption (Zhou et al., 2018). Optimization model with multi- travel to access their workplace, go to the market to access some goods &
objective particle swarm optimization algorithm also showed better services, and for going back to their homes. The purpose of travel is not
performances in terms of reduction in the average passenger waiting traveling but to access something in time & space. That’s why the de­
time by 10.54% and greater vehicle occupancy ratio (Hartmann Tolić mand for transportation is considered as derived demand. There are two
et al., 2020). In the case study of Tehran Metropolitan rail network, essential dimensions of urban transportation demand.
adaptive particle swarm algorithms successfully reduced the total pas­
senger waiting times significantly about 64% and 9.4% better perfor­ Time
mance compared to the current practice of using an even-headway The demand for urban transportation services varies with the time of
timetable as well as the baseline timetable (Hassannayebi et al., 2018). the day, that is measured in terms of hourly passenger load from origin
Metro train timetable also depends upon various elements in its to destination. The demand is highest in morning hours (0700–1030
optimization such as factor affecting arrival-departure schedule from hrs.) when commuters have to reach their offices and evening hours
stations following headway, peak & off-peak variable headway as per (1630–2000 hrs.) when commuters leave their office for their homes.
passenger flow (Lei et al., 2016), coaches utilization as per passenger This kind of variation in the demand as per the time of the day is termed
demand, train & station capacity to accommodate passengers safely for a as a temporal variation of demand which can be observed from Fig. 1,
comfortable ride (Guo et al., 2016), multiple loops in operation to depicting the number of passengers traveling from one place to another
maintain minimum headway in most crowed section, run time between in an hour or hourly variation in the demand in terms of hourly Peak
source to terminal as per speed regulation coefficient, reversal plan at Hour Peak Direction Traffic (PHPDT) for line-2 of Delhi Metro Rail Cor­
terminals to decrease train reversal time (Marinov et al., 2014), train poration (DMRC) [Jahangirpuri (JGPI) to Huda City Centre (HCC)] for a
identity dedicated for 6 or 8 coach train, piloting train movement to typical day.
check track and structure after maintenance and before the start of
revenue service, train movement after last revenue service over rusty Space
points and crossings which are not in normal service use, details of first The demand for urban transportation services also depends upon the
& last revenue to display as passenger information, revenue & non- place. The commercial places, residential places, central business dis­
revenue trains (Horz, 2016). Revival of demand responsive transport tricts, educational hubs are major traffic generators whereas demand in
(DRT) service in some areas helped urban planners and operators in thinly populated areas of the city is less. This kind of variation in the
their decision-making process. DRT service comparison on various fac­ demand for urban transport as per the characteristics of the place is
tors to a regular fixed service, based on an actual case study in termed as spatial variation in demand which can be observed from
Amsterdam showed that the reduced mileage and operating timeframe Fig. 2, depicting the spatial variation along line-2 of DMRC during
contributed to the better overall efficiency of the on-demand system, morning peak hours (0730–1000 hrs.) in terms of entries plus exits at
despite the drop in the ridership (Coutinho et al., 2020). different stations for a typical day.
Some literatures discuss the demand & supply equilibrium of trans­
portation services based on Timetable rationalization & Operational Supply of urban transport services
performance improvements (Pan et al., 2020; Xu et al., 2020; Scheep­
maker et al., 2020) whereas, some research focused on the timetable The supply of transportation services is measured in terms of the
optimization of urban transit systems in aspect of travel time reduction availability of commuter accommodation in the vehicle from one place
only (Chang and Chung, 2005; Shang et al., 2016), but this work ex­ to another during any hour of the day. If the number of trains is operated
amines the human intervention expressing demand & supply equilib­ per hour from source to terminal at a particular time with defined train
rium with cost optimization. This study is successful in considering capacity, then the supply of the services is equal to the product of the
different operational strategies with conditions illustrated as: number of trains and defined train capacity.

• Optimizing coach composition, metro train capacity, train frequency Methodology

and dwell time with dynamic timetable is considered to cater vari­
ation in sectional ridership during peak & off-peak hours following The objective of any research primarily involves steps carried out in
passenger demand and high capacity utilization; the study as demonstrated in Fig. 3. Here in step-1, the field of study is
• Multiple loops operation in the most crowded section is considered to discussed where the research study depending upon resources & its
introduce peak & off-peak variable headways as per passenger flow constraints is carried out. In step-2 timetable rationalization method is
and to maintain minimum headway precisely in peak hours; discussed with various types of timetables. In step-3 amelioration of
• Adjustment in metro train speed during off-peak hours is considered timetable is carried out in contrast to various factors responsible for
to ensure safety with energy efficiency during weather or infra­ timetable amelioration. In step-4 results are presented after methodol­
structure operating conditions like turnout, horizontal and vertical ogy adopted in step-2&3, followed by constraints of the study, and
gradients and platform entry; lastly, the conclusion is discussed in step-5.
• Change in reversal plan at terminals is considered to decrease train An urban rail transit system is amongst the most critical and valuable
reversal time as the braking pattern and sectional clearance are transportation infrastructure for metropolitan cities to perform suc­
related to train speed; cessfully on the global competitive node. Manpower and Energy costs
• Net energy consumption is considered as the end result impacting are the key components of operational expenditure in an urban rail
operational expenditures. transit system. The substantial segments of energy cost are traction and
auxiliary consumption. In traction, train operation and HVAC (heating,

2
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 1. Hourly demand for line-2 for a typical working day.

Fig. 2. Spatial variation in demand for line-2.

ventilation and air-conditioning) are interdependent and play major during communication or Passenger Information Display System & Pas­
roles. The traction consumption in an urban rail transit system is the end senger Announcement System (PIDS & PAS) failures, they should have a
result of operational strategies with efficient usage of timetable & assets. fair idea of the destination of train originating from a particular station
An efficient timetabling ensures not to indulge excessive trains in or at least aware of the maximum time they may have to wait at the
operation and takes care of bottlenecks during intermediate reversals. interchange to get the train for the desired direction.
Planned headway should be based upon the actual time taken during This study considers only revenue commuters as they are related to
reversals at the terminals and should have a cushion to accommodate an urban transit system for direct revenue generation and avoiding all kinds
unprecedented delay in service, minimum to the extent of two minutes of incidents causing delays and an increase in the system running cost.
in one direction. Special train movements like piloting train and rusty For timetable rationalization supply must meet the demand and this
movement should be well defined. The need-based speed profile, if strategy is proposed in this study with an objective of scheduling train
permitted by the system to save energy, shall be possible for improving timings from the origin to the terminus, Operational changes in the
the energy efficiency of train operations, e.g., maximum speeds during timetable, and real-time data monitoring to maintain equilibrium be­
peak hours and coasting mode during off-peak when people are gener­ tween supply & demand as per station ridership and Automatic Fare
ally not in a hurry to reach their destination. It should not be too Collection (AFC) PHPDT. In this research study, the goal is highlighted as
complicated to create confusion in the minds of the passengers, even timetable rationalization & operational improvements with human

3
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

considered, as this is the industrious operational line in the entire

network of DMRC in terms of passenger flow, Fig. 4(b).
This study tried to fully utilize all the resources available only within
DMRC system such as AFC data to calculate timely entry & exit from the
system, ridership data to calculate passenger using metro services from
first revenue to last revenue service, trainload to calculate passengers
count per train, train & coach capacity in respect of gender-based pas­
senger flow like the male, female, physically challenged for seat & coach
reservations considering safety, run time & dwell time for timetable
rationalization & service improvements, and lastly the ICONIS archi­
tecture interface by ALSTOM with full passenger capacity revenue ser­
vices and without passenger non-revenue services in all weather
conditions.

Timetable rationalization

The type of timetable to be used largely depends upon the passenger

flow or demand profile, hourly as well as geographical variations in
Fig. 3. Research methodology. demand. The passenger flow data is collected from the AFC system to
extract AFC data, ridership data, and PHPDT of 34 serving DMRC Sta­
intervention to meet up the rising passenger demand in form of ridership tions. It is computed as the sum of Contactless Smart Card Exit, Con­
and lowering the operational cost in terms of energy consumption. tactless Smart Token Entry, Free Exit Tokens, Paid Exit Tokens, Tourist
Card Exit, Staff Pass, Employee Stored Value Card Exit, Group Tickets,
and VIP visits hourly throughout the day. Here some of the data is
Field of study collected by AFC systems (like Contactless Smart Card Exit + Contactless
Smart Token Entry + Free Exit Tokens + Paid Exit Tokens + Tourist
Delhi has well established urban railway network connecting Delhi- Card Exit + Employee Stored Value Card Exit), while rest of the data
National Capital Region with 11 operational lines covering 391 km & (like Staff Pass + Group Ticket + VIP visits) is collected manually and
285 serving stations, Fig. 4(a). But in this research study of timetable digitize by Revenue department to combine with AFC data for utilization
rationalization, only line-2, yellow line of DMRC (49 km stretch & 37 and passenger flow evaluation. Generally, in metro systems following
serving stations from Samaypur-Badli (SPBI) to HCC of which 3 station types of timetables are being used:
SPBI, Rohini Sector-18 (RISE) & Badli-Haiderpur-Mor (BIMR) has not
been considered due to very low ridership & timetable requirement) End to end operation
covering 45 km & 34 serving stations (JGPI to HCC) has been If the demand remains significant across the line, the trains are

Fig. 4. (a). Route map of Delhi Metro network, (b). Line-2 (Yellow line) of DMRC. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)

4
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

planned for an end-to-end operation (JGPI-HCC-JGPI loop) without any low demand. Since the supply is being matched with the demand, this
intermediate reversal point. Here end to end terminology describes both kind of timetable leads to the optimal utilization of assets. However, the
ends of the line-2/yellow line in terms of loop operation like JGPI to level of service to the commuters is slightly lower than that in constant
HCC to JGPI train movement, which is considered as a single loop headway timetable as waiting time in the lean periods is higher causing
operation. In this type of timetable, all trains originate from one ter­ little inconvenience to the passengers.
minal (JGPI) and operate up to another terminal (HCC) and vice versa, In DMRC, the timetable is fixed for months & changes are considered
thereby completing a loop run, except for depot going trains which as per ridership requirements only where demand for urban transport
originate/terminate at the station serving toward the depot. This is the services is frequently changing variable (hourly/daily/monthly) across
simplest kind of timetable from an operational point of view of the time and space. Demand has an impact on timetable designing but it
system as well as commuters as depicted in Fig. 5(a). cannot be premixed as a variable in the timetable. Demand & its vari­
ation is considered in both cases during designing & rationalization of
Constant headway between trains timetable but tabulation data of hourly demand during a month is a
If demand throughout the day does not much vary, the supply should quiet bit lengthy to represent in consideration of keeping the theoretical
also remain constant i.e. the same number of trains should run aspect of study under permissible limits. During the study, various
throughout the day at a constant headway as shown in Fig. 5(b). This fabricated timetables were loaded into the DMRC signaling system and
kind of timetable offers a high level of service to the passengers as train found that end to end timetable and constant headway timetable were
services are available at a constant headway at any time of the day. not enough to support the purpose of study due to high operational
However, the temporal variation in urban transport demand is rather expenditure in terms of energy consumption, while intermediate
normal than an exception. The constant headway timetable results in reversal timetable and variable headway timetable alone was not sup­
suboptimal utilization of assets in periods of low demand. porting the purpose to cater the PHPDT. Thus, most optimal timetabling
is achieved by adopting variable headway and intermediate reversals
Intermediate reversals combination with different dwell time as per demand variation in AFC
Generally, the demand at the extreme ends of a line is significantly data, ridership, and PHPDT.
lower than that in the central portion of the line. End to end operations
in such cases would result in the supply of services much more than the
Amelioration of timetable
demand in extreme ends of the line and consequent suboptimal utili­
zation of the resources. Initially, there is no intermediate reversal
Analysis of demand and supply
planned in the timetable, because Scheduled mode as shown in Fig. 5(b)
DMRC is an interconnected network of 11 different lines, where the
is not operational but during timetable rationalization, intermediate
direction is way more complicated. But for the sake of simplicity, this
reversal is introduced to stop end to an end operation. In intermediate
study is based only on the operational aspects of the line-2/Yellow line
reversal timetabling, there may be one or more intermediate reversal
of DMRC where JGPI to HCC section is considered to highlight only two
points on the mainline from where trains are reversed regularly to meet
directions of travel (JGPI to HCC/ Down line and HCC to JGPI/ Up line).
the higher demand in the central portion of the line fulfilling minimum
Regarding interconnectivity there are only two interchanging in line-2,
headway. These are complex and should be supported by proper inter­
Kashmiri Gate (KG) & Rajeev Chowk (RCK) from where footfall data is
locking design at an intermediate reversal station with the passenger
collected from the Origin-Destination matrix & the same is combined
information system.
with AFC data for utilization. This study has utilized the hourly AFC data
of line-2 in combination with footfall from interchanging stations
Variable headway between trains
instead of sufficient detailing about complete network and considering
In this kind of timetable, the supply varies as per the temporal
demand per origin–destination, to keep the theoretical aspect of study
variation of demand. The headway of the trains is variable i.e., small
under permissible limit. The demand & supply data aggregated over a
headway at the time of higher demand and large headway at the time of
time interval of an hour for different sections of the lines have been

Fig. 5. (a). End to End operation timetable; (b). Constant headway timetable.

5
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

obtained against the time of the day. The PHPDT data for the demand of offered in the evening peak (1730–2000 hrs.) by matching AFC
the services has been obtained from the AFC system and found that PHPDT data.
PHPDT during the evening peak (1730–2000 hrs.) is 14.57% less than • Reduction in dwell time at the stations for increasing the average
that in the morning peak (0730–1000 hrs.), so there is a chance of speed of the train including different dwell times at different stations
pushing fewer trains in the evening peak as compared to morning peak as per the station ridership.
hours. The supply of the services has been computed based on the • Utilization of front crossover at the JGPI, VW, and QM for train
number of trains being operated per hour as per the prevailing time­ reversal instead of siding to save reversal time.
table. The demand & supply curves for different sections of line-2 of • Deployment of an extra train operator at the source & terminal
DMRC are illustrated in Figs. 6(a) to (f) and the following observations (JGPI, VW, QM, HCC) for immediate reversal & departure.
were made:
Similar measures have been taken on other lines also. The number of
• The demand at the ends of the line [Qutubminar (QM) to HCC and trains operated in different lines on morning peak, off-peak, and evening
Vishwavidyalaya (VW) to JGPI] is almost 53.68% that of VW-QM peak with demand and supply is presented in Table 1.
section in peak hours in terms of passengers per hour per direction.
• In off-peak hours (1100 to 1700 hrs.), the supply of accommodation Real-time monitoring of demand
(number of trains) is far more than the demand (PHPDT) resulting in Any timetable for a metro system is developed for a given pattern of
low occupancy of trains during off-peak hours and consequent sub- the demand considering spatial and temporal variation in the demand.
optimal utilization of resources. However, there may be significant variations in the pattern of the de­
mand itself due to the day-specific events like nearby cultural events/
Measures for matching demand & supply function, political/religious rally, exam center, incidents, and protests in
This study has tried to rationalize the timetable by way of taking the city. The static timetable would not be capable to absorb such sud­
supply-side measures i.e., meeting the demand with an adequate num­ den variations in demand and would result in inconvenience to pas­
ber of trains (supply) and ensuring an acceptable level of services. With sengers in case demand is more than supply or inefficiency in the system
the availability of peak & off-peak hour demands, the train capacity is if demand is significantly lower than the supply. This study tried to
provided. Predefined per coach capacity based on the train configura­ devise a method to monitor the demand on a real-time basis with the
tion of 6/8 coach composition declares single train capacity, that can help of entry & exit at the stations from 0500 hrs. to 1100 hrs. (6 h
determine exactly the number of trains required to meet the demand. To continuously) and compare it with the entry & exit data of an average of
rationalize the timetable of line-2 in DMRC as illustrated in Fig. 7 the last 4 weeks for the same time frame & same day of the week. This
following measures have been taken: comparison gives a fair idea of the pattern of the demand for the day to
regulate train requirement in afternoon (non-peak hours) & evening
• Reversal of 1/3rd trains from VW and QM during morning and eve­ (peak hours). It is evidently supporting the passenger demand future
ning peak hours. projections to induct or withdraw the trains for the next 6 h (1100 hrs. to
• Reversal of alternate train at VW and QM during off-peak hours. 1700 hrs.) may be taken well in advance. This method majorly focuses
• Nearly 33% reduction in the number of trains from 53 earlier to 36 on token entry having a predefined zone of travel as per the token sale
now being offered in off-peak hours (1100–1700 hrs.) and 8% amount and card exit because of the undefined zone of travel of
reduction in the number of trains from 53 earlier to 49 now being commuter. In real-time monitoring method hourly AFC data is
computed to add/remove the trains from mainline due to sudden surge

Fig. 6a. Comparison of line capacity & AFC PHPDT in line-2 JGPI-VW DN.

6
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 6b. Comparison of line capacity and AFC PHPDT in line-2 VW-QM DN.

Fig. 6c. Comparison of line capacity and AFC PHPDT in line-2 QM-HCC DN.

in commuter entry & exit from the metro system. Herewith an increase withdrawal of one or few trains from mainline to depot or preferably
in ridership, induction of more trains from depot or preferably sidings sidings performed as per requirements.
for single or multiple loops run and with a decrease in ridership Real-time monitoring method:

7
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 6d. Comparison of line capacity and AFC PHPDT in line-2 HCC-QM UP.

Fig. 6e. Comparison of line capacity and AFC PHPDT in line-2 QM-VW UP.

• World bank study on urban & rail-based transportation, indicates transit capacity increases by 15% to match the demand and
to increase PHPDT by 10% in any section/line, Either reduce the supply exactly by 10%, still urban transit has 5% surplus capacity
fare to attract passenger toward public transport use or increase for comfortable journey of passengers.
the capacity (supply) by 10% to attract and cater more PHPDT as • With reference to above strategy, a reference data showing day-
service enhancement by drawing passenger feedback. If urban wise average entry & exit up to 1100 hrs. with ± 10% variation

8
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 6f. Comparison of line capacity and AFC PHPDT in line-2 VW-JGPI UP.

Fig. 7. Timetable rationalization of the line-2/Yellow line of DMRC.

Table 1
Demand & Supply scenario of trains at different intervals of time.
Line No. of Trains Capacity PHPDT

Morning Peak Off-Peak Evening Peak Morning Peak Off-Peak Evening Peak Morning Peak Off-Peak Evening Peak

1 24 17 24 22,040 14,500 21,460 19,027 9372 17,208

2 53 36 49 58,000 34,800 45,240 47,022 34,220 40,167
3&4 61 50 60 45,820 32,480 42,340 43,721 18,827 33,335
5 15 13 15 13,468 11,396 13,986 10,028 4718 7055
6 23 11 19 19,684 9324 14,504 14,759 6342 13,046

and whole day entry & exit has been prepared for the last 4 i. If on given day number of entry & exit (up to 1100 hrs.) is up to
weeks. This data is dynamic which is updated weekly i.e., the +/- 10% of the average number of entry & exit (up to 1100 hrs.)
average figures are updated for the last 4 weeks at the end of the of reference table, the train shall continue to run as per the
week which is used for the upcoming week. Any exceptional day timetable.
is not included while calculating the average. ii. If on a given day the number of entry & exit (up to 1100 hrs.) is
• AFC provides line-wise data (no. of entry & exit up to 1100 hrs.) decreased by more than 10% of the average number of entry &
by 1130 hrs on daily basis to the chief controller/Operation exit (up to 1100 hrs.) as per reference table, the number of trains
Control Center. will be reduced by the actual percentage change in the number of
• The chief controller analyzes this data and suggests the reduction entry & exit (up to 1100 hrs.) adjusted by 10% of the buffer. e.g.,
of trains as per the following guidelines: if current day number of entry & exit (up to 1100 hrs.) is reduced

9
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

by 15% of the average number of entry & exit (up to 1100 hrs.) of Table 2
reference table, the number of trains for rest of the day will be Additional train induction in the peak of peak hours.
reduced by 5% (15% − 10% = 5%). Line Period Section No. of additional No. of trains
iii. The withdrawn trains are sent to depot or terminals or preferably trains induction withdrawal in
sidings. in peak hours off-peak hours
iv. In line-2, preferably 8 coach trains are withdrawn if required. 2 Morning Vishwavidyalaya to 1 18
Similarly in line-3&4, preferably 6 coach trains are withdrawn if Peak Sultanpur
required. Evening Sultanpur to 1 14
Peak Vishwavidyalaya
• Exceptions: 3&4 Morning Inderprastha to 1 12
i. In case the next day is a holiday or a big event, this procedure is Peak Dwarka
not implemented. Evening Dwarka to 1 11
ii. If there was any failure during the first half (up to 1100 hrs.) of Peak Inderprastha
6 Morning Sarita vihar to 1 7
the day then this exercise is not carried out as AFC entry & exit
Peak Central secretariat
figures up to 1100 hrs. might not represent actual no. of pas­ Morning Central secretariat 1 7
sengers in the system. Peak to Badarpur
iii. During failures or incidents affecting train operation after 1100 Evening Sarita vihar to 2 6
hrs., the number of trains in operation will be governed by the Peak Central secretariat
Evening Central secretariat 2 6
requirement to deal with the impact of failure and not with the Peak to Sarita vihar
percentage of average entry & exit up to 1100 hrs.

Sequencing of 6 & 8 coach trains in line-2 these days is even less than that on Sundays. Hitherto, DMRC was
In an urban rail transit system project length of metro train depends running trains as per Sunday timetable on these days. Now, the time­
upon the length of the platform and sidings. The ridership in an urban table for these days has been further rationalized as per the estimated
transit system takes time for building up and based on the ridership demand.
projections, platforms in DMRC are designed for a maximum of 8 car
capacity train, but only 4 car capacity trains were utilized during initial Special timetable for weekends during festival season
operation to cater the PHPDT available and to limit capital expenditure. It has been observed that the ridership during festive weeks for
Train capacity from 4 to 6 cars and 6 to 8 cars is evolved in due course of example at the time of Holi, Diwali, Eid, Christmas, Dussehra, Onam,
time of ridership escalation. Pongal, Muharram, and Vinayaka Chaturthi remains comparatively low
Presently, line-2 operates trains of 6 & 8 coach compositions i.e., as compared to other weekends. Special timetables are being planned for
mixed fleet usage. The 6 coach trains are being progressively converted weekends during the festival weeks.
into 8 coach trains to augment carrying capacity. Due to large scale
conversion in fleet size to meet up the escalated demand in ridership Coasting mode in off-peak hours
with the same train frequency & time interval, same headway, single The traction energy consumption in the train operations largely de­
train operator, the system is bound to utilize mixed fleet, but 6 coach pends upon factors like trainload, the efficiency of traction motor, speed
trains cannot withstand the crowd in peak timing. Thus, the intention is profile of the train, acceleration & deceleration, operating skills of the
to utilize 8 coach trains in peak hours in high ridership VW-QM section train operator, regenerative braking, air conditioning load, etc. The
and 6 coach trains in low ridership sections such as JGPI-VW & QM-HCC maximum traction energy is drawn during motoring and there is
as an objective of rationalization of timetable. The dynamic timetable is wastage of traction energy in braking as all mechanical energy can’t be
designed with a strong focus on earmarked identities of 8 car trains only converted into electrical energy through regenerative braking. The
for high ridership section as an operational criterion that none of them traction energy consumption can be significantly optimized if a train is
should enter in low ridership section in peak hours and only 8 car trains driven in coasting mode (constant speed) for most of its travel. The ideal
should be withdrawn from service in off-peak hours. The train depots are speed profile of a train would entail maximum acceleration while
also advised to ensure induction of the 8 coach trains strictly as per starting from a station to achieve maximum speed in minimum time,
earmarked identities. With the relentless support of the train depot, an cruising and coasting for the middle part of the journey followed by
adequate supply of 8 coach trains in the most crowded section during maximum deceleration while approaching the station to stop the train at
peak hours is ensured. Thus, instead of running 8 car trains whole day, the right location. Coasting and Energy-saving speed profiles in Auto­
staggered timing for 8 car train operation is opted and 6 car trains are matic Train Operation (ATO) mode watch for energy saving in zero &
utilized whole day to maintain a balance between energy consumption negative gradients midsection where the train can move with its mo­
and timetable optimization. mentum in the midsection between two stations. Thus, trains do not
need to apply motoring/braking, consume energy and easy rolling is
Induction of additional trains at the peak of peak hours followed in zero gradients midsection and coasting followed by braking
The demand during peak hours is also variable to some extent. This in negative gradient midsection, which also supports energy saving in
study considered the demand variation in peak hours and assessed peak operation. The speed profile for maximum coasting and energy-saving is
demand during peak hours which is termed as a peak of peak demand. presented below in Fig. 8.
Concerted efforts were made to meet the peak of the peak demand by However, this kind of speed profile is very difficult to achieve in
inducting additional trains in the different section of line 2, 3, 4, and 6. manual train operations as speed profile would vary according to the
The additional train induction in the peak of peak hours is done by operating skills of the train operator. In ATO mode, the motoring,
experimental method on human experience and AFC ridership data braking, and stoppage of the train are controlled by the system. The ATO
basis. For train regulation, withdrawal in off-peak hours as discussed in mode of train operation offers six kind of train speed profiles truly
table-2 is done by steps mentioned in real-time monitoring method dependent on cruising & coasting performed in the mid-section, as
explained in para 2.3.3; Table 2 presented in Table 3. More the coasting more is the energy saving on the
verge of more time consumption to commute from one place to another.
Special timetables for National/Gazetted holidays and festivals Here for the convenience of the passengers, the run time should be
This study considered the ridership on national/gazetted holidays minimal which remains so, if speed profile changes from Energy Saving
and festivals in the last three years and observed that the ridership on to Tight Run, due to which our coasting in midsection tends to decrease

10
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 8. Typical coasting & Energy saving mode speed profile.

Table 3
Speed profiles in ATO mode with Energy consumption by train in line-2.
Speed Speed Regulation JGPI-HCC Runtime (in minutes) Coasting Energy
Profiles Coefficient (assuming dwell time of 25 sec) Consumption

Tight Run 100% 00:67:52 Minimum Maximum

Normal 80% 00:69:37
Tight
Normal 60% 00:71:01 To To
Running
Normal Un- 40% 00:72:12
tight
Coasting 20% 00:72:27
Energy 0% 00:73:27 Maximum Minimum
Saving

Energy Consumption in kWh/km

Train I.D. Train Set Period Tight Running Coasting Energy Saving

JGPI-HCC/ HCC- JGPI- HCC-JGPI/Up JGPI- HCC-

Dn JGPI/Up HCC/Dn HCC/Dn JGPI/Up

8 Coach 215 BT-07 Peak 13.545 12.727 11.093 11.041 10.227 10.021
218 BT-12 Off- 11.318 11.068 10.698 10.443 9.860 9.841
Peak
6 Coach 225 BT-39 Peak 11.295 10.636 10.773 10.372 10.326 10.302
232 BT-36 Off- 10.841 10.659 10.144 9.364 9.159 9.093
Peak

on the verge of increase in energy consumption. However, to curtail station.

surplus energy consumption per kilometer as specified in Table 3, the
trains should run maximal either in Energy-saving or Coasting mode, Results
which also entails an increase in run time. The timetable needs to bal­
ance the two conflicting requirements of the passengers and the oper­ By following the steps mentioned in para 2.0, during the study, 220
ator. During the study, DMRC has tried to run the trains in ATO mode trains were withdrawn from service as per the real-time assessment of
with ‘Tight Run/Normal Tight’ speed profile during peak hours to save demand and results in terms of comparison of energy consumption (per
time of commuters to reach their destination and ’Coasting’ profile in day) before & after the adoption of timetable rationalization for line-2
off-peak hours to save cost in terms of energy consumption by train for a half month is presented in Table 4.
which follow coasting on various track circuits calculated by onboard Evidently, energy consumption has decreased initially by only 0.26%
ATS signalling system. Here in coasting profile onboard ATS provides due to the operation of more 8 coach trains only as a conditional non-
only 5 to 8% motoring after achieving desired speed to cope up the revenue trial run of 500 km/train before induction into passenger rev­
gradients during run. enue service. In January 2018 total testing run during non-revenue
This methodology ensures the maximum level of service to the pas­ hours was 38,355 km. In addition to this ridership is also increased by
sengers during peak hours and a reasonable level of service during off- 19.17%, affecting trainload as well as energy consumption. But a major
peak hours. Further, to make up time lost in coasting mode opera­ decline of 2.10% in energy consumption in April 2018 clearly depicts
tions, available dwell time at the stations is reduced. The reduced dwell that the rationalization method performed well in energy saving, and
time has not caused any inconvenience to the passengers as with pro­ potent support of the measures adopted. Further, the pre-rationalization
gressive replacement of 6 coach trains with 8 coach trains, the number of and post-rationalization demand and supply curves for line-2 are also
entry/exit doors per train has increased by 8 (4 per coach) which means illustrated below in Figs. 9 & 10 for appreciating the matching of de­
the lesser time required for boarding/de-boarding (dwell time) at any mand and supply in the rationalized timetable.

11
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Table 4
Energy consumption before & after the adoption of timetable rationalization for line-2.
Traction energy consumption before timetable Traction energy consumption after timetable rationalization
rationalization

Date Traction Energy Consumption (kWh) Date Traction energy consumption (kWh) Date Traction energy consumption (kWh)
With Mixed Fleet Usage of 6 Coach Train = 12 With Mixed Fleet Usage of 6 Coach Train = 9 With Mixed Fleet Usage of 6 Coach Train = 7
8 Coach Train = 41 8 Coach Train = 44 8 Coach Train = 46

01.12.17 385,220 01.02.18 387,650 01.04.18 382,119

02.12.17 374,574 02.02.18 341,920 02.04.18 343,141
03.12.17 399,640 03.02.18 400,660 03.04.18 391,529
04.12.17 412,115 04.02.18 419,620 04.04.18 394,022
05.12.17 406,330 05.02.18 398,780 05.04.18 391,708
06.12.17 413,945 06.02.18 411,580 06.04.18 408,004
07.12.17 419,587 07.02.18 429,190 07.04.18 414,296
08.12.17 409,110 08.02.18 413,290 08.04.18 407,499
09.12.17 371,972 09.02.18 310,510 09.04.18 304,819
10.12.17 346,823 10.02.18 360,470 10.04.18 350,253
11.12.17 353,369 11.02.18 366,310 11.04.18 355,999
12.12.17 365,618 12.02.18 364,560 12.04.18 356,751
13.12.17 365,480 13.02.18 372,310 13.04.18 373,972
14.12.17 385,800 14.02.18 413,610 14.04.18 411,437
15.12.17 389,026 15.02.18 392,730 15.04.18 391,129
Total 5,798,609 5,783,190 5,676,678
Energy Saving Units (kWh) 15,419 121,931
Energy Saving Percentage 0.26% 2.10%

Fig. 9. Demand Vs Capacity before Timetable Rationalization.

With the results of the timetable rationalization approach imple­ • Alteration of dwell time on a real-time situation basis on the platform
mented in line-2/yellow line of DMRC, including the ATO profile with is not feasible irrespective of the passenger boarding/de-boarding
Coasting/Energy saving, certain constraints which are encountered time and time required for a physically challenged person.
during the study listed as: • ATO profile has only front cab dependency, if due to any reason the
front cab becomes inactive/delocalize/faulty then the train can’t
• If there is a faulty train ahead inside the tunnel or in the middle of the move in ATO, It has to depart in ATP mode following rear cab
overhead section, ALSTOM ATO operation of the previous train communication which indirectly causes a delay in service due to
could not be able to stop the train at the platform. It allows the train complete human-based operation.
to move up to a safer distance behind faulty train inside tunnel or in • Train following timetable fed into the system continues to apply
the middle of the overhead section irrespective of time required for motoring & braking only to maintain target speed profile, it does not
fault clearance. follow coasting mode for energy saving on its own as a signaling
feature, while a human can be judgmental as per operating
constraints.

12
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Fig. 10. Demand Vs Capacity after Timetable Rationalization.

Conclusion Kumar Chauhan: Visualization, Supervision, Writing – review &

editing.
Before implementation of the measures, unmatched demand &
supply was resulting in constant headway in more & less crowded sec­
tion, unnecessary operational trips, Inter-depot rakes transfer for major Declaration of Competing Interest
maintenance, overhauling and testing, which further results in energy
consumption & wastage. This study has analysed the operational per­ The authors declare that they have no known competing financial
formance of DMRC based upon changing needs of operations by devising interests or personal relationships that could have appeared to influence
& implementing experimental measures in timetable rationalization to the work reported in this paper.
strike a balance between ensuring a comfortable level of service and
optimal utilization of resources.
References
After timetable rationalization as per methodology adopted during
the operational years 2017–2018, it reveals optimum train operation in Aksu, D.T., Akyol, U., 2014. Transit coordination using integer-ratio headways. IEEE
terms of saving in energy consumption and operational expenditure, Intell. Transp. Syst. Mag. 15 (4), 1633–1642. https://doi.org/10.1109/
TITS.2014.2301821. [19].
thus materializing the cost factor as low as reasonably practicable. Based
Albrecht, T., 2009. Automated timetable design for demand-oriented service on
on the results It can be concluded that in order of timetable ration­ suburban railways. Public Transp. 1 (1), 5–20. https://doi.org/10.1007/s12469-
alization & operational performance improvement this study is suc­ 008-0003-4.
cessfully matching supply with demand across the day by applying: Barrena, E., Canca, D., Coelho, L.C., Laporte, G., 2014. Single-line rail rapid transit
timetabling under dynamic passenger demand. Transp. Res. Part B: Methodol. 70,
134–150. https://doi.org/10.1016/j.trb.2014.08.013.
• Variable headway with an off-peak coasting mode of operations in Boroun, M., Ramezani, S., Vasheghani Farahani, N., Hassannayebi, E., Abolmaali, S.,
ATO for energy saving; Shakibayifar, M., 2020. An efficient heuristic method for joint optimization of train
scheduling and stop planning on double-track railway systems. INFOR: Inform. Syst.
• Real-time monitoring of passenger demand to optimize the fleet Operational Res. 58 (4), 652–679. https://doi.org/10.1080/
requirement in operations; 03155986.2020.1746100.
• Interlacing of high-capacity trains during peak hours helps in uti­ Caprara, A., Fischetti, M., Toth, P., 2002. Modeling and solving the train timetabling
problem. Oper. Res. 50 (5), 851–861. https://doi.org/10.1287/opre.50.5.851.362.
lizing resources more judiciously; and, Carey, M., Crawford, I., 2007. Scheduling trains on a network of busy complex stations.
• Performing intermediate reversals, which helps adequate passenger Transp. Res. Part B: Methodol. 41 (2), 159–178. https://doi.org/10.1016/j.
clearance at crowded sections and attaining better headway. trb.2006.02.002.
Castillo, E., Gallego, I., Urena, J.M., Coronado, J.M., 2009. Timetabling optimization of a
single railway track line with sensitivity analysis. TOP 17 (2), 256–287. https://doi.
All the measures described in this work are adopted as per the need org/10.1007/s11750-008-0057-0.
and it can also be used as a benchmark by other metros with their own Castillo, E., Gallego, I., Urena, J.M., Coronado, J.M., 2011. Time-tabling optimization of
a mixed double and single-tracked railway network. Appl. Math. Model. 35 (2),
experience which are experiencing similar challenges in their
859–878. https://doi.org/10.1016/j.apm.2010.07.041.
operations. Castillo, E., Gallego, I., Sánchez-Cambronero, S., Menéndez, J.M., Rivas, A., Nogal, M.,
Grande, Z., 2015. An alternate double-single track proposal for high-speed
peripheral railway lines. Comput.-Aided Civ. Infrastruct. Eng. 30 (3), 181–201.
CRediT authorship contribution statement Ceder, A., 1984. Bus frequency determination using passenger count data. Transp. Res.
Part A: General 18 (5–6), 439–453. https://doi.org/10.1016/0191-2607(84)90019-
0.
Manish Kumar Sharma: Conceptualization, Methodology, Valida­ Ceder, A., 2001. Bus timetables with even passenger loads as opposed to even headways.
tion, Formal analysis, Investigation, Writing – original draft. Bhavesh Transp. Res. Rec. 1760 (1), 3–9. https://doi.org/10.3141/1760-01.

13
M. Kumar Sharma and B. Kumar Chauhan Transportation Research Interdisciplinary Perspectives 13 (2022) 100526

Ceder, A., 2001. Efficient timetabling and Vehicle scheduling for public transport. In: Liebchen, C., 2008. The first optimized railway timetable in practice. Transp. Sci. 42 (4),
Vol, S., Daduna, J.R. (Eds.), Computer-Aided Scheduling of Public Transport, 505. 420–435. https://doi.org/10.1287/trsc.1080.0240.
Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-56423-9_3. Marinov, M., Giovanni, L.D., Bellisai, G., Clevermann, J., Mastellou, A., Victόria, D.,
Ceder, A., 2007. Public transit planning and operation: Theory, Modeling and Practice. Deleva, L., 2014. Analysis of rail yard and terminal performances. J. Transp.
Elsevier, Butterworth-Heinemann, Oxford, United Kingdom. Literature 8 (2), 178–200. https://doi.org/10.1590/S2238-10312014000200008.
Chang, S.-C., Chung, Y.-C., 2005. From timetabling to train regulation, a new train Niu, H., Zhou, X., 2013. Optimizing urban rail timetable under time-dependent demand
operation model. Inf. Softw. Technol. 47 (9), 575–585. and oversaturated conditions. Transp. Res. Part C: Emerg. Technol. 36, 212–230.
Chen, Q., 2014. Global Optimization for Bus Line Timetable Setting Problem. Discrete https://doi.org/10.1016/j.trc.2013.08.016.
Dyn. Nat. Soc. 2014, 1–9. Palma, A.D., Lindsey, R., 2001. Optimal timetables for public transportation. Transp.
Coutinho, F.M., van Oort, N., Christoforou, Z., Alonso-González, M.J., Cats, O., Res. Part B: Methodol. 35 (8), 789–813. https://doi.org/10.1016/S0191-2615(00)
Hoogendoorn, S., 2020. Impacts of replacing a fixed public transport line by a 00023-0.
demand responsive transport system: Case study of a rural area in Amsterdam. Res. Pan, D., Chen, Z.J., Mei, M., 2020. Energy efficiency emergence of high-speed train
Transp. Econ. 83, 100910. https://doi.org/10.1016/j.retrec.2020.100910. operation and systematic solutions for energy efficiency improvement. SN. Applied
Cury, J., Gomide, F., Mendes, M., 1980. A methodology for generation of optimal Sciences 2(5), article no, 875. https://doi.org/10.1007/s42452-020-2692-5.
schedules for an underground railway system. IEEE Trans. Autom. Control 25 (2), Robenek, T., Maknoon, Y., Azadeh, S.S., Chen, J., Bierlaire, M., 2016. Passenger centric
217–222. https://doi.org/10.1109/TAC.1980.1102282. train timetabling problem. Transp. Res. Part B: Methodol. 89, 107–126. https://doi.
Feng, X., Mao, B., Feng, X., Feng, J., 2011. Study on the maximum operation speeds of org/10.1016/j.trb.2016.04.003.
metro trains for energy saving as well as transport efficiency improvement. Energy Scheepmaker, G.M., Willeboordse, H.Y., Hoogenraad, J.H., Luijt, R.S., Goverde, R.M.P.,
36 (11), 6577–6582. https://doi.org/10.1016/j.energy.2011.09.004. 2020. Comparing train driving strategies on multiple key performance indicators.
Feng, X., Liu, H., Zhang, H., Wang, B., Sun, Q., 2013. Rational Formations of a Metro J. Rail Transp. Plann. Manage. 13, 100163. https://doi.org/10.1016/j.
Train Improve Its Efficiencies of Both Traction Energy Utilization and Passenger jrtpm.2019.100163.
Transport. Math. Probl. Eng. 2013, 1–7. Shang, P., Li, R., Yang, L., 2016. Optimization of Urban Single-Line Metro Timetable for
Furth, P.G., Wilson, N.H.M., 1981. Setting frequencies on bus routes: theory and practice. Total Passenger Travel Time under Dynamic Passenger Demand. Procedia Eng. Issue
Transp. Res. Rec. 818, 1–7. 137, 151–160.
Guo, B., Zhou, L., Yue, Y., Tang, J., 2016. A Study on the practical carrying capacity of Sun, L., Jin, J.G., Lee, D.H., Axhausen, K.W., Erath, A., 2014. Demand-driven timetable
large high-speed railway stations considering train set utilization. Math. Probl. Eng. design for metro services. Transp. Res. Part C: Emerg. Technol. 46, 284–299. https://
Article ID- 2016, 1–11. https://doi.org/10.1155/2016/2741479. doi.org/10.1016/j.trc.2014.06.003.
Hartmann Tolić, I., Nyarko, E.K., Ceder, A.A., 2020. Optimization of Public Transport Taboada, G.L., Han, L., 2020. Exploratory Data Analysis and Data Envelopment Analysis
Services to Minimize Passengers’ Waiting Times and Maximize Vehicles’ Occupancy of Urban Rail Transit. Electronics 9 (8), 1270. https://doi.org/10.3390/
Ratios. Electronics 9 (2), 360. https://doi.org/10.3390/electronics9020360. electronics9081270.
Hassannayebi, E., Sajedinejad, A., Mardani, S., 2014. Urban rail transit planning using a Wang, Y., Schutter, B.D., Boom, T.V.D., Ning, B., Tang, T., 2013. Real-time scheduling for
two-stage simulation-based optimization approach. Simul. Model. Pract. Theory 49, single lines in urban rail transit systems 108, 1–6. https://doi.org/10.1109/
151–166. https://doi.org/10.1016/j.simpat.2014.09.004. ICIRT.2013.6696258.
Hassannayebi, E., Zegordi, S.H., Amin-Naseri, M.R., Yaghini, M., 2018. Optimizing Wong, R.C.W., Yuen, T.W.Y., Fung, K.W., Leung, J.M.Y., 2008. Optimizing timetable
headways for urban rail transit services using adaptive particle swarm algorithms. synchronization for rail mass transit. Transp. Sci. 42 (1), 57–69. https://doi.org/
Public Transp. 10 (1), 23–62. https://doi.org/10.1007/s12469-016-0147-6. 10.1287/trsc.1070.0200.
Horz, P., 2016. Remembering railway’s past, conjuring up its future: what rail hikers Wu, C., Pei, Y., Gao, J., 2015. Model for Estimation Urban Transportation Supply-
have in mind while walking on rusty tracks, in: McCants, A., Beira, E., Cordeiro, J.M. Demand Ratio. Math. Probl. Eng. 2015, 1–12.
L., Lourenco, P.B., Pereira, H.S., New Uses for Old Railways. TUA, pp. 15. Xu, Y., Jia, B., Li, X.P., Li, M.H., Ghiasi, A., 2020. An integrated micro-macro approach
Koutsopoulos, H.N., Odoni, A., Wilson, N.H.M., 1985. Determination of headways as a for high-speed railway energy-efficient timetabling problem. Transp. Res. Part C-
function of time varying characteristics on a transit network. Comput. Scheduling Emerg. Technol. 112, 88–115. https://doi.org/10.1016/j.tre.2020.01.008.
Public Transp. 2, 391–413. Zhou, Y., Bai, Y., Li, J., Mao, B., Li, T., 2018. Integrated Optimization on Train Control
Lei, D., Guo, C., Zhang, Y., Chen, C., 2016. Reasonable scheduling for arrival–departure and Timetable to Minimize Net Energy Consumption of Metro Lines. J. Adv. Transp.
track operations in railway stations. Transp. Plann. Technol. 39 (6), 624–639. 2018, 1–19.
https://doi.org/10.1080/03081060.2016.1187812.

14