Ways to Improve the Efficiency of the

Truck's Branded Service System

K. Shubenkova

, P. Buyvol

, I. Makarova

and L. Gabsalikhova

Kazan Federal University, 68/19 Mira Avenue, Naberezhnye Chelny, Russia

Keywords: Branded Service System, Dealer Service Center, Simulation Model, Statistical Data Analysis.

Abstract: The article discusses ways to improve the trucks maintenance efficiency. It is shown that only integrated

solutions will optimize the activities of the automotive corporation's branded service system (BSS). The best

solution in this situation is a decision support system (DSS) with an open architecture. The proposed method-

ology is aimed at improving the maintenance and repair system while expanding markets. Examples of de-

veloped modules applying as part of DSS, such as statistical data analysis and simulation models, are shown.

1 INTRODUCTION

The economy globalization, as well as the rapid

development of engineering and technology and

increased competition in the markets, have shortened

the launch time for new goods. High-tech products

require service during the entire life cycle, so should

ensure trouble-free operation, which implies the

responsibility of the manufacturer to the client. In the

automotive industry, this concept is implemented by

creating a branded service system (BSS). As a rule, the

system includes a network of dealer & service centers

(DSC). Regardless of format, such DSC operate in

accordance with the manufacturer standards. Two

aspects must be considered. The first is the vehicle's

maintenances & repair (M&R) systems quality, which

should keep the vehicle in working condition. The

second is the process quality of providing services to

the client - the vehicle owner. This is important

because manufacturer competitiveness and brand trust

depend on the BSS effectiveness. It is necessary to

correctly evaluate the performance services indicators

and risks, as well as provide actions that will help to

avoid risks or will be needed in risk situations case.

The most common way to ensure accessibility is to

create a reserve. These actions can be divided into two

directions. The main reserve capacities are formed on

https://orcid.org/0000-0002-9246-6232

https://orcid.org/0000-0002-5241-215X

https://orcid.org/0000-0002-6184-9900

https://orcid.org/0000-0003-3325-3285

the basis of demand forecasting and provide estimated

performance. This helps to insure against errors in

forecasts and from possible delays in the current orders

execution. The second direction of service efficiency

increase is realization of the customer-oriented

approach. Ensuring customer loyalty should be

considered as a prerequisite for achieving the success

of the company in the competition (Lovelock et al.,

2011).

One way to increase the competitiveness of both

the entire BSS and each of the DSCs is to regulate the

processes in each of them according to manufacturer

reviews. Currently, there are tools for working with big

data, so many vehicles manufacturers create tools for

collecting and processing large data amounts that

connect all production, logistics and service enterprises

into a single information space. This allows you to

control each vehicle throughout the entire life cycle

and helps to optimize all processes. At the same time,

it is possible to analyze the work results, compare them

with previous periods and give recommendations for

further adjustment of the development strategy. To

make adequate and justified decisions, Decision

Support Systems (DSS) are developed, the main

module of which is the intelligent core, which is

responsible for obtaining optimal decisions. This

approach is especially relevant in situations where

Shubenkova, K., Buyvol, P., Makarova, I. and Gabsalikhova, L.

Ways to Improve the Efﬁciency of the Truck’s Branded Service System.

DOI: 10.5220/0009839206730680

In Proceedings of the 6th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2020), pages 673-680

ISBN: 978-989-758-419-0

673

resources are limited, or in the case of the new vehicle

models launch on the market (Buyvol et al., 2019).

2 PROBLEM STATUS: BRANDED

SERVICE SYSTEM

2.1 Simulation Models in DSS as a Way

to Find Optimal Solutions

Upon entering new markets, vehicle manufacturers

create BSS abroad. As a rule, this is a network of DSC

authorized in accordance with the manufacturer's

standards. In most cases, they organize their activities

on the principle of “three S”: sale - service - spare parts.

For trucks, this approach is most relevant, since the

share of trucks in the total fleets number is rather small,

and it is more difficult to organize a maintenance in

small workshops than to cars. In addition, the

maintenance cost, as well as the complexity of

servicing trucks is higher (Makarova et al., 2013;

2015). Decision Support Systems (DSS) contain three

main subsystems: 1) module for data collection and

storage that comes from internal and external sources

(usually this is a data warehouse); 2) module for data

processing and analysis - the intellectual core; 3) the

user interface, which is necessary for interaction and

communication of clients using information flows.

This allows you to select data for analysis and

parameters that affect the management decision. The

DSS conceptual diagram is shown in Fig. 1. Simulation

models as part of the intellectual core provide the

search for the best solutions in various activity fields

of the entire service system and each of its subsystems.

Information from a common data warehouse in which

data is constantly updated is used to determine model

parameters. Updating the parameters allows you to

find the best solution in each specific situation.

The main problem here is the quality, completeness

and adequacy of statistics' data. Since the reliability of

complex technical systems depends on many factors,

the analysis quality will depend on the various data

groups quality. An incorrect situation interpretation

due to incorrect source data can significantly affect the

frequency of failures. Because of this, the information

quality about the vehicle technical condition at the

failure time and the conditions of its operation

preceding the failure allows the manufacturer to

improve not only the vehicle design, but also the

warranty service system. Statistical information is used

not only to create a simulation model, but also to verify

its adequacy and compliance with the real system

(validation and verification). When trucks M&R

planning during the warranty period, the information

quality problem and model adequacy can become

critical (Mikulec et al., 2017, Srinivasana et al., 2016).

This is due to the increased failures number in the

initial operation period, and the manufacturer must

remove the failure consequences in accordance with

the warranty service contract. In many cases, it is

related to use incomplete or subjective information,

which is contained in the complaint acts drawn up

when the vehicles owners contact the DCS. To obtain

more correct information, data is often used from an

intelligent vehicle on-board system. In the research

(Meeker et al., 2014), it is shown that various sensors

installed on the product can be used to collect

information about how, when, at what environmental

parameters and under what conditions the product is

operated. This approach is suggested in the paper (Last

et al., 2010). They state that it is possible to use multi-

target algorithm of estimated probability for a

probability prediction and a choice of failure time in

system of guarantee maintenance. For reliability

modeling authors use Weibull’s analysis.

2.2 Processes Organization in the

Branded Service System

Diagnosing faults in automotive systems is an

important stage in M & R, as it affects the duration to

complete the work. For fault taxonomy, a fault tree

diagram is often used. However, since the system's

structure is implicit, the article (James et al., 2018)

authors propose digraph modelling method, which

uses the graph theory's system approach. The proposed

approach contains recommendations for diagnosing

the main failures causes. Methodology

computerization will help in creating a knowledge base

about failures and how to resolve them. Therefore, this

approach is especially useful for M & R engineers.

Vehicle Health Management (VHM) often

includes real-time monitoring of operating conditions,

as well as decision-making on driving, operating, and

maintenance based on predicted conditions. The article

(Jaw et al., 2004) presents a universal, flexible

integration and testing concept for control evaluating,

including the accuracy of decision-making, algorithms

and models in real time and in closed cycle.

Meckel S. (Meckel et al., 2019) propose methods

for extracting knowledge from unstructured and

informal materials on Internet forums, offering more

effective and targeted actions for diagnostics and

maintenance in real time. This is necessary for the

synthesis of diagnostic graphs from the created

knowledge base for use in vehicle maintenance.

iMLTrans 2020 - Special Session on Intelligent Mobility, Logistics and Transport

674

Figure 1: DSS conceptual scheme.

The study (Börger et al., 2019) goal is to reduce the

time required for trucks maintenance. It can be done by

applying the Lean methodology. The article (Vintr et

al., 2003) goal is to find ways to optimize the concept

of maintenance for reduce the life cycle cost (LCC) of

a vehicle based on operational reliability data. The

authors indicate that it is relatively easy to find reserves

and achieve significant savings in the vehicle LCC

using a simple administrative measure of change in the

maintenance frequency.

In order to maintain a high vehicle operability level

and the transport system safety as a whole, it is

necessary to adhere to a strategy and an appropriate

schedule for vehicle maintenance. In the paper (Kamlu

et al., 2019), to develop a fuzzy model, a condition-

based maintenance strategy is proposed that takes into

account various types of uncertainties for individual

vehicles, such as, for example, load, mileage and

terrain using either wired or wireless data and to

failure's predicts.

3 CASE STUDY: THE USE OF

MODELING TO SOLVE

PROBLEMS OF PROCESS

OPTIMIZATION IN THE BSS

3.1 Application of DSS in Strategic

Planning in the Corporate Service

System

Statistical data in the BSS can be used in the following

ways. Firstly, on the data array basis for a certain pe-

riod, it is possible to establish the main DCS activity

parameters (the requests flow intensity or services, the

requests number, the average services laboriousness)

and use them to perform an optimization experiment

and improve processes. Secondly, an analysis of the

failure distributions allows us to identify the problems

causes and find a way to eliminate them.

Thirdly, the failures flow trend analysis makes it

possible to predict the failures number in the

prospective periods of DSC activity and improve the

spare parts supplies management. The distribution

laws parameters analysis consists in comparing the

data set characteristics different periods and

developing methods for making managerial decisions.

In this case, as the optimal distribution in the current

period, a distribution law is selected that has the

estimates values (variance and mathematical

expectation) closest to the real data. Such laws are built

for each detail, unit and aggregate in a special DSS

software module “Statistical data analysis for DSC”.

A change in the distribution law parameters

indicates to change in the analysing indicator. For

example, if the optimal parameters among the

distribution laws in previous periods are the parameters

M = 16650 (km) And  = 177560 (km

), and as a result

of the current data sample analysis, the parameters M

=14350 (km) and  =213430 (km

), this indicates that

the average mileage of the vehicle to failure for

investigated vehicle's component has decreased, while

the random values spread has increased, which is most

likely caused by a quality decrease to this component's

manufacturing.

The user interface form of the module “Statistical

Data Analysis for DCS” has two tab cards: “Prediction

of component failures” and “Monitoring of service

parameters” (fig. 2). When you select the “Vehicle

mileage statistics” option, a selective two-dimensional

data set is formed according to two parameters:

“component name” and “vehicle mileage”, and then is

exported to the Statistica software application.

Ways to Improve the Efﬁciency of the Truck’s Branded Service System

675

Based on this set, for a given component, a

frequency distribution histogram of the range values is

built, after which the distribution law that best

describes the resulting array is selected. As Fig. 3

shows the vehicle mileage value distribution laws until

the failure of the “ST142-10 starter” component on

KAMAZ 4320 truck model, according to data for the

2018 first quarter. The decision-making method

consists in choosing the distribution law with the best

agreement criterion and comparing this law parameter

of with the previous periods parameters. For this, the

data is transferred to a special decision-making module

developed in MS Excel (Fig. 4), in which compiles a

summary table of distribution laws parameters.

Figure 2: DSS statistical analysis module.

Figure 3: The distribution laws of the mileage to failure.

Figure 4: Decision-making module for DCS management.

Figure 5: Comparison of gamma-distribution law parame-

ters for the four quarters of 2018.

Thus, it can be seen from the presented example

that the random vehicle mileage to defective

component failure conform the gamma distribution

law, while the law parameters change in the direction

optimal for the enterprise (Fig. 5), that is, the shape

parameter increases, which affects the mathematical

expectation of the random magnitude, and also

decreases the scale parameter, affecting the random

variable's variance. In the general case, it can be

predicted that over time, the gamma distribution law

will be transformed into the normal distribution law.

The algorithm for the module “Forecasting the

failures number” allows you to create a time series of

the replacements number of the defective component

for the period specified by the user, which is also

exported to the Statistica software, where the model of

the seasonal component is selected (additive or

multiplicative), and a time series line and an

extrapolated trend line with a relative less than 10%

error (Fig. 6). The tool “Monitoring service

parameters” allows, on the basis of calls to the service

center, to determine the DCS activity parameters

during the period specified by the user. The data sets

corresponding to the selected service parameters are

transferred to the Statistica software (Data Science …,

2020), where they are processed statistically. The

module window view is shown in Fig. 7.

Figure 6: The problem solution to the component failures

numbers predicting.

iMLTrans 2020 - Special Session on Intelligent Mobility, Logistics and Transport

676

Figure 7: Select service monitoring options.

Figure 8: Service Time's Distribution Laws (hours).

Figure 9: Distribution Laws of interval between request re-

ceipt’s (hours).

In fig. 8 shows the distribution laws of the service

execution time value in hours (for all database defects).

From the above graphs it follows that the Weibull

distribution law with the parameters  = 28.76 and  =

1.29 is the most suitable, since it is the only one that

has a significance level p = 0.10125 that exceeds the

specified level  = 0.1. Figure 9 shows the distribution

law of the random interval value between the service

requirements receipt in hours.

From the above graphs it follows that the most

suitable is the exponential distribution law with the

parameter  = 1.035, since it is the only one that has a

significance level p = 0.13401 that exceeds the

specified level  = 0.1. The calculations result of the

queuing system parameters are used to develop a

simulation model and conduct an optimization

experiment. The analysis results serve to develop and

adjust instructions intended for both service centers

and vehicle owners. Compliance with these

instructions can improve the vehicle operation

reliability.

Constant monitoring of the service system status

allows you to increase not only the vehicle reliability,

but also the DSC efficiency. If the distribution laws

parameters indicate a deterioration in the system state

(for example, a decrease in the average vehicle mileage

to failure), then the management in the previous step

was not rational, and the system needs a control action.

In this case, recommendations for the control action are

developed on the optimization experiment basis a

conducted on a simulation model. In the case when the

parameters values of the distribution law indicate an

improvement in the parameters of the system

functioning, we can conclude that the control in the

previous period was rational, that is, adequately by set

goals (Khabibullin et al., 2013).

3.2 M & R Processes Simulation in

DSC

When new vehicle models appear, it is necessary to

evaluate the capabilities of the existing dealer and ser-

vice network (DSN), which can be done using simu-

lation models. After simulation model creating and

checking its adequacy, it can be used for an optimiza-

tion computer experiment, why allows you to find a

control impact in which performance indicators will

be optimal for the system under certain external con-

ditions. When creating the model, two approaches

were combined: agent modeling (agents - vehicles)

and discrete event modeling (service process execu-

tion in the DCS). This made it possible to combine the

queuing system principles with the stochastic behav-

ior model of separate objects.

Figure 10: Agent-based model for «vehicle» object.

The agent “vehicle” can be in two states: “service-

able” and “repair required”.

The transition from the first state to the second is de-

termined by the vehicle's failure probability function

to mileage. The time to return to working condition

(“The average time of fault removal” - Y) is deter-

mined by the parameters: the percentage of requests

for spare parts directly from the warehouse (X1), the

number of workstations in the DSC (X2), the number

of workers per workstation (X3), the distribution of

Ways to Improve the Efﬁciency of the Truck’s Branded Service System

677

arrivals from the vehicles concentration point in the

DSC (X4), the schedule of the DSC (X5) (Fig. 10).

In addition, when generating “vehicle” for each

agent, the “type” property is determined (base vehicle,

truck tractor, trailer vehicle, dump truck, specialized

vehicle), which affects the repair work duration in ac-

cordance with the standards. To develop a simulation

model, we used the library of discrete event modeling

objects (Enterprise Library) in the AnyLogic software

application (Borshchev, 2014). Besause trucks in most

cases represent corporate fleets, two classes were cre-

ated during the simulation: Auto (agent class, models

one truck) and VCP (models the vehicle concentration

place). When the model is started, for each “Auto”

agent, the initial mileage value (which then constantly

increases by the flowAuxVar value), the maximum

vehicle mileage and the mileage to failure are gener-

ated. Upon reaching the mileage to failure, the agent

state goes from working (“Serviceable”) to Ou-

tOfOrder (“Repair is required”), a need for M&R is

formed, and vehicle is delivered to DSC for services.

After the vehicle is transportation to DSC, the ser-

vice algorithm corresponding to the created class

agent “Service station” is applied to it. So, if all the

parking spaces for vehicles awaiting repair are occu-

pied, the service request is rejected and the vehicle

leaves the DSC. After the repair station is vacated, the

availability necessary spare parts for repair are

checked. If there is no necessary spare part, the request

arrives at the delay unit simulating the spare parts ex-

pectation, after which the request falls into the “de-

ferred repair” block and leaves the service system. To

verify the proposed approach adequacy, we chose the

Kazakhstan Republic (RK) DSN, which has 16 DSS.

This market is actual for KAMAZ in competition with

Chinese manufacturers conditions, so improving the

efficiency of this BSS segment is important. Since the

RK territory has four climatic zones with different op-

erating conditions (Köppen, 2011), the law on the fail-

ure distribution was specific for each of them. Crite-

rion function of system management model estab-

lishes balance between of manufacturer investments

on DSN development and the lost profit from client’s

loss in view of admissible queue length excess:

- Z

 max (1)

where 𝑍



– benefit of additional clients' service at the

taken measures of DSN development,

= E - KL

inv

(2)

E – average benefit of one served client, rub;

inv

–

the difference between quantity of served clients be-

fore and after development; Z

– costs of investments

of DSN development,

= P + N

inv



(3)

P – expenses on information support, rub; N

inv

- num-

ber of added DSC in DSN; S

– cost of additional DSC

building rub.

Full system effectiveness is defined by decrease in

client's losses that depend on excess of vehicle delay

time in BSA in comparison with a specified time on

declared works implementation, and also DSC ex-

pense minimization depending on equipment and

worker’s downtimes. The client’s losses connected

with delivery and vehicles delay in the DSC:

∑























𝑇



1 









𝑇





𝑆



𝑁



→𝑚𝑖𝑛







where:𝑆



– the average client’s losses due to with ve-

hicle shutdown, rub/hour; 𝑁



– number of vehicles,

server in 𝑗 DSC; 𝐷



–average vehicle transportation

distance to the 𝑗 DSC, km; 𝑉 – vehicle transportation

speed to the DSC, km/h; 𝑅 – the number of DSC; 𝑇



– standard delivery time on spare parts to j DSC,

hours; 𝑇





– vehicle repairs average time in 𝑗 DSC,

hours; 𝑋





– number of workers per one station in 𝑗

DSC; 𝑇





–average expectation time in service queue

in 𝑗 DSC, hours; 𝑋





– satisfaction percent for spare

parts demands directly from a warehouse in 𝑗 DSC.

DSC expenses connected with shutdowns:

∑

𝑆



∙𝑋





𝑆



∙𝑋





∙𝑋





∙𝑇





→𝑚𝑖𝑛





(5)

where: 𝑆



– costs associated with stopping one work-

station on hour (missed profit), rub/hour; 𝑆



– average

salary per hour, rub/hour; 𝑇





– average time stopping

one workstation in j DSC, hours; 𝑋





–number of work-

station in j DSC.

Model's limitations:

1. Limitation on exceeding the investment

amount of economic benefits

> 0 (6)

2. Limitation on the maximum investments

amount that the manufacturer is ready to invest for

progressing for the DSN development:

>INV (7)

3. 𝑋





, 𝑋





– limitation on the warehouse

space size for storage of the minimum and maximum

spare parts volume:

4. 𝑋





𝑋





𝑋





(8)

∑

𝑁









∑∑

𝑋





∙









𝑉𝐶𝑃



(9)

iMLTrans 2020 - Special Session on Intelligent Mobility, Logistics and Transport

678

𝑋





– distribution of arrivals from 𝑖-place of

fleet concentration to 𝑗 DSC, %,

𝑉𝐶𝑃



– number of inoperative vehicles in 𝑖-

place of park concentration.

6. 𝐾





, 𝐾





- coefficients of the min-

imum and maximum admissible workload of work-

stations in j-DSC.

7. 𝐾











∙











∙





∙∙





𝐾





(10)

where 𝐷 – quantity of days in simulated period; 𝑇





– the average time of vehicle repairs in j- district, hour;

𝑇





– working shift duration in j- district, hours; 𝑋





–

number of working shifts in j-district, hours.

8. 𝑋





, 𝑋





– minimum and maximum

normative workstations number of 𝑗 DSC (limit the

stations number).

𝑋





𝑋





𝑋









9. 𝑋





, 𝑋





– minimum and maximum

workers quantity on workstations in 𝑗 DSC (limit on

the labor resources number).

𝑋





𝑋





𝑋









The model was verified by the tracing method

(Sargent, 2011). Since each DSC is a queuing system

with a specified number of parallel workstations, to

verify the model, the average workload on the work-

station was compared with the calculated use coeffi-

cient for the selected time period 𝜌



𝑚∙𝑡







𝑛∙

𝑡





, where 𝑚 – vehicles quantity being repaired funds,

𝑡



– average repair time of one vehicle, 𝑛 – quantity

of workstations in DSC, 𝑡



– workstation capacity for

the considered period (Introduction to …, 2008). Dur-

ing the simulation experiment, data were used on the

fleet's species-age structure and the DSC characteris-

tics. As the system's response vector, we used the av-

erage time spent by the vehicle in the DSC, belonging

to the corresponding format's group, which were ob-

tained as a clustering result. Clustering was carried out

according to the estimated parameters by the k-means

method based on dendrograms based on project and

calculated parameters. The graph of cluster averaging

over estimated parameters is shown in Fig. 11.

The developed simulation model's adequacy was

evaluated according to the statistical theory of

assessment and hypothesis testing, using the following

criteria:

1. Dispersions of the model’s responses

deviations from the average values of systems

response. Dispersions comparison was performed to

Fisher criterion. The results, presented in Tab. 1, show

that in all three clusters 𝐹𝐹



, i.e. the hypothesis of

the differences importance between the two variance

estimates is rejected.

2. Using the Student t-test, we tested the

hypothesis that the average values of each n-

component of the Y

model responses are close to the

average values of the n-component of the real system

𝑌′



responses. For the real system and simulation

model, the expected value and dispersion 𝑌′



, 𝐷′



and

𝑌



, 𝐷



, were estimated (Tab. 2) (Buyvol et al., 2019).

The calculation results show that for all three clusters

𝑡



𝑡



, i.e. the hypothesis on proximity of the

responses average values for model and the system is

accepted.

Figure 11: Graph of a clustering averaging by estimated pa-

rameters (1 - complaints number; 2 - warranty vehicles num-

ber; 3 - services volume; 4 - sold spare parts cost; 5 - sold

vehicles cost; 6 - sold vehicles number (units); 7 - operations

results for the reporting period (profit/loss); 8 - profitability

of sales; 9 - services profitability; 10 - fulfillment of cus-

tomer service requirements.

Table 1: The results of comparing dispersions according to

the Fisher test.

№

system γ1 D

F F

Real system 3 2.247

3.68 4.76

Simulation model 6 0.610

Real system 3 0.868

0.47 8.94

Simulation model 6 1.864

Real system 3 2.648

1.93 8.94

Simulation model 6 5.100

Table 2: The hypothesis verification results using t-student

test.

№ system N Y

∆ω

Real system 4 37.05

1.16 1.409 2.262 0.78

Simulation model 7 36.75

Real system 4 46.34

1.53 1.80 2.26 1.41

Simulation model 7 47.00

Real system 4 41.90

4.28 0.21 2.26 1.39

Simulation model 7 42.48

4 CONCLUSIONS

Scientific attitude at BSS improvement helps to react

to the arising problems at new model vehicles

Ways to Improve the Efﬁciency of the Truck’s Branded Service System

679

operation quickly, having provided possibility of it

constructions improvement. As the executed

researches have shown that only system solutions for

increasing the vehicle reliability at all life cycle stages

will make it possible to increase its safety, as well as to

ensure the possibility of trouble-free operation. The

decision-support systems for management

improvement use will allow to correct the actions,

which directed on strategic goal realization at each

stage. Statistical data analysis and simulation

modelling as the intelligent block main element of DSS

will allow selecting the most rational variant for each

real condition combination. At the same time, it is

necessary to create conditions for initial data timely

updating, its operative processing and ready solutions

storage.

ACKNOWLEDGEMENTS

This work was supported by the Russian Foundation

for Basic Research: grant No. 19-29-06008 \ 19

REFERENCES

Börger, A., Alfaro, J., León, P., 2019. Use of the Lean Meth-

odology to Reduce Truck Repair Time: A Case Study.

In: Rocha Á., Adeli H., Reis L., Costanzo S. (eds) New

Knowledge in Information Systems and Technologies.

WorldCIST'19 2019. Advances in Intelligent Systems

and Computing, vol 930. pp 655-665.

Borshchev, A., 2014 Multi-method modelling: AnyLogic, in

Discrete-Event Simulation and System Dynamics for

Management Decision Making. John Wiley & Sons Ltd.

Chichester, U.

Buyvol, P. et al., 2019. Forecasting of Changes in Service

System During the Launch Period of the New Automo-

bile Lineup. Helix. Vol. 9 (4): 5221- 5226.

Data Science Textbook. 2020. URL: https://docs.tibco.com/

data-science/textbook.

Introduction to Discrete-Event Simulation. Chapter 10.

2008. In: Cassandras C.G., Lafortune S. (eds) Introduc-

tion to Discrete Event Systems. Springer, Boston, MA.

pp. 557-615.

James, A.T., Gandhi, O.P. & Deshmukh, S.G. 2018. Fault

diagnosis of automobile systems using fault tree based

on digraph modeling. Int J Syst Assur Eng Manag 9,

494–508.

Jaw, L. and Wang, W. 2004. A run-time test system for ma-

turing intelligent system vehicle capabilities - SIDAL,

2004 IEEE Aerospace Conference Proceedings, Big

Sky, MT. Vol.6. pp. 3756-3763.

Kamlu, S., Laxmi, V., 2019. Condition-based maintenance

strategy for vehicles using hidden Markov models. Ad-

vances in Mechanical Engineering. Vol. 11(1) 1–13.

Khabibullin, R.G. et al., 2013. The study and management

of reliability parameters for automotive equipment using

simulation modeling. Life Science Journal. 10 (12s),

132, pp. 828-831.

Köppen, W., 2011. The thermal zones of the Earth according

to the duration of hot, moderate and cold periods and to

the impact of heat on the organic world. Meteorolo-

gische Zeitschrift, Vol. 20, No. 3, 351-360.

Last, M., Sinaiski, A., Subramania, H.S. 2010. Predictive

Maintenance with Multi-Target Classification Models.

Intelligent Information and Database Systems. Lecture

Notes in Computer Science. 5991: 368-377.

Lovelock, C.H., Wirtz, J. 2011. Services Marketing: People,

Technology, Strategy, 7th Edition. Published by Pren-

tice Hall. 612 p.

Makarova, I. et al. 2013. Improving of performance system

of warranty for automotive engineering abroad on the

basis of data of rejections analysis. Innovative Mechan-

ical Engineering Technologies, Equipment and Materi-

als-2013 Vol. 69.

Makarova, I. et al. 2015. Improving the system of warranty

service of trucks in foreign markets. Transport Prob-

lems. Vol. 10. Iss.1 pp. 63-78.

Meckel, S. et al., 2019. Optimized Automotive Fault-Diag-

nosis based on Knowledge Extraction from Web Re-

sources, 24th IEEE International Conference on Emerg-

ing Technologies and Factory Automation (ETFA), Za-

ragoza, Spain, pp. 1261-1264.

Meeker, W.Q., Hong, Y. 2014. Reliability Meets Big Data:

Opportunities and Challenges. Quality Engineering, 26.

pp.102–116.

Mikulec, N., Felke, T., Bangale, S. 2017. Analysis of War-

ranty Data to Identify Improvements to Vehicle Relia-

bility and Service Information. SAE Int. J. Passeng.

Cars – Electron. Electr. Syst. 10(2).

Sargent, R.G., 2011. Verification and validation of simula-

tion models. WSC '11 Proceedings of the Winter Simu-

lation Conference. Phoenix, Arizona, pp. 183-198

Srinivasana, R. et al. 2016. Modelling an Optimized War-

ranty Analysis methodology for fleet industry using data

mining clustering methodologies with Fraud detection

mechanism using pattern recognition on hybrid analytic

approach. Procedia Computer Science, 87, pp. 322 –

327.

Vintr, S.Z. and Holub R., 2003. Preventive maintenance op-

timization on the basis of operating data analysis. An-

nual Reliability and Maintainability Symposium, 2003.,

Tampa, FL, USA. pp. 400-405.

iMLTrans 2020 - Special Session on Intelligent Mobility, Logistics and Transport

680