=Paper=
{{Paper
|id=Vol-2843/shortpaper43
|storemode=property
|title=Intellectualization of decision making in information systems lifecycle management in fintech (short paper)
|pdfUrl=https://ceur-ws.org/Vol-2843/shortpaper043.pdf
|volume=Vol-2843
|authors=Alexander Kvasov,Yuri Yegorov,Vladimir Milov
}}
==Intellectualization of decision making in information systems lifecycle management in fintech (short paper)==
https://ceur-ws.org/Vol-2843/shortpaper043.pdf
Intellectualization of decision making in information
systems lifecycle management in fintech*
Alexander Kvasov, Yuri Yegorov and Vladimir Milov
Nizhny Novgorod State Technical University n.a. R.Е. Alekseev, Nizhny Novgorod, Russia
ak@2cio.ru
Abstract. The organizational aspects and technologies for modeling the proc-
esses of information system life cycle management in fintech are considered.
The current growth rates of the amount of information requiring processing in
real time, as well as the trend towards the creation of an increasing number of
remote jobs in various sectors of the economy, force us to reconsider the views
on the organization of the information systems lifecycle management process.
Keywords: Intellectualization of decisions, Information systems, Technologies
for modeling.
1 Introduction
The relevance of creating methods for describing and modeling information flows is
caused by a high degree of intensification of information exchange and the prevalence
of the use of new information technologies. The construction of models of life cycle
(LC) management processes of information systems (IS) is designed to provide sup-
port to decision makers (DM) for timely intervention in the development of the in-
formation system of the company.
Timely identification of trends that can negatively affect the life cycle of a finan-
cial organization IS, their elimination through its revision and updating will not only
adequately calculate the necessary resource costs for system support, but can also
serve as a signal for the need to make changes in the established business processes of
the organization [1].
Despite the fact that, in general, the IS includes almost all of the organization's re-
sources that one way or another process information or participate in information
flows, the most important is that part of the financial organization's IS that provides
automation of the main functions of business units. It is usually called an automated
banking system or corporate information system (CIS). It is the CIS that is considered
by the authors as an object of research, the modeling of the life cycle of which can be
extrapolated to the IS of fintech.
*
Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribu-
tion 4.0 International (CC BY 4.0).
The life cycle model of the corporate information system of a financial organiza-
tion is a combination of a sequence of life cycle stages and transitions between them,
necessary to ensure the achievement of the goal set for the project. The specificity of
the life cycle of fintech CIS is that the initial stages of the life cycle, such as the sur-
vey of the subject area, the statement of the problem, design and development, now
no longer make sense, since very few new financial organizations are opening, but for
them there are many ready-made automated systems on the market, the implementa-
tion of which significantly reduces the period for launching a CIS.
Deployment and implementation for fintech are very complex processes, since they
take a significant amount of time, while the business must remain continuous. The
likelihood of the need to dispose of the current corporate information system and the
introduction of a new one is intended to minimize for fintech this study in an effort to
preserve the current corporate information system. We will focus on the stages of
Operation, support and modernization of the corporate information system. The deci-
sions made on them should ensure maximum compliance of corporate information
systems with business requirements at minimum costs.
The operation of the corporate information system, which is the main stage, the
balance of which must be maintained, is ensured by the IT department of the organi-
zation through the stability of the operation of hardware, system software and end-
user support. At this stage, the intellectualization of the decisions made in the man-
agement of the life cycle of the corporate information system is especially important,
since it helps prevent serious consequences associated with a decrease in the qualita-
tive and quantitative indicators of the corporate information system in comparison
with the reference ones that correspond to the changing business requirements [8].
2 Statement of the corporate information system support and
modernization
2.1 Criteria for assessing the compliance of a corporate information system
with the requirements of a business customer
In their activities, employees of the organization are guided by certain regulations that
describe business processes that can be assessed in terms of the time spent, and hence
the level of automation [2]. Subsequent monitoring of the compliance of this assess-
ment with the average or reference value can be an important indicator of the need to
make changes both in the corporate information system and in the business process
itself. Multi-criteria assessment and monitoring of the main business processes indica-
tors will allow avoiding critical situations when the revision of the corporate informa-
tion system may be either economically ineffective or ineffective in terms of meeting
business requirements.
In addition to the time indicator, the degree of automation and formalization of
tasks performed by employees is considered. A decrease in this indicator entails addi-
tional costs in time for manual processing of documents.
The convenience of the CIS interface, and, consequently, the laboriousness of em-
ployees performing their business functions is the most difficult to perceive factor,
however, significantly affecting both the speed of information processing and the
efficiency of the organization as a whole. Routine actions are no longer critically
assessed by performers, but newly hired employees can draw attention to the low
level of business processes automation. At the same time, the analysis and control of
such situations can be automated.
The launch of the CIS revision and adaptation processes to the rapidly changing
business requirements is possible only through constant monitoring of many factors
and processes occurring in the corporate information system. However, for this it is
necessary to formalize these factors and processes as much as possible.
2.2 Application of the apparatus of fuzzy logic in assessing the relevance of
CIS
The use of fuzzy logic makes it possible to assess the relevance of the current CIS
and to select an adequate organizational impact on its life cycle. If the corporate in-
formation system does not provide the required efficiency of the business process, its
revision may be required (setting the configuration of the interconnected components
that make up the corporate information system). In a critical case, the introduction of
a new IS may be required [9].
It is necessary to choose such a conceptual solution (configuration) that, with
maximum utility (or quality) of the IS, ensures the efficiency of the business process,
while meeting the requirements is a necessary condition, but does not guarantee the
achievement of maximum efficiency (the ratio of effectiveness and costs).
Usefulness - the degree of user satisfaction with a set of functions (modules) of the
CIS. Since the revision of the CIS is characterized by a given labor intensity and risks
(design or technical), the effectiveness can be assessed using a composite criterion
that includes indicators such as Ui – the utility of the i-th component of the CIS (for
example, a module that generally performs one function) , Ti is the complexity of the
implementation of the CIS component (set during business planning) and Pi is the
probability of the successful implementation of the СIS component (estimated by
experts).
The search for systemic and technical solutions, including the optimization of the
U, T, P indicators of individual components, will maximize the usefulness of the cor-
porate information system, which will help improve the efficiency of the business
process.
However, the high rates of development of information technologies lead to the
fact that CIS requires constant improvement. On the information technology market,
there are many new practices aimed at active interaction of IT professionals, both
developers and administrators, in order to increase the pace of their interaction and
improve the efficiency of issuing updates to the CIS. For example, DevOps (devel-
opment & operations), supplemented by a flexible approach to Agile development,
focused on the use of iterative development, dynamic formation of requirements and
ensuring their implementation as a result of constant interaction within self-
organizing working groups, consisting of specialists of various profiles.
The DevOps approach forms the continuous interaction of employees in the IT de-
partment, without which the benefits of life cycle modeling and the intellectualization
of decisions made can be reduced to zero. If, for example, administrators do not have
time to deploy the releases supplied by the developer into production, updates accu-
mulate, and business customers do not receive the necessary functionality.
In addition, recently the so-called microservices (MS) have become popular, on the
basis of which the CIS of a financial organization is built. MS are used as building
blocks to automate various parts of business processes [3]. Microservices architecture
has its advantages in upgrading and scalability of СIS, but it also contains disadvan-
tages associated with managing the growing number of microservices.
Modern information technologies that make it possible to bring the process of up-
dating and upgrading the corporate information system to a completely new quality
level using adaptive algorithms and using knowledge bases significantly reduce the
costs and terms of implementing improvements to the corporate information system
and minimize the risks associated with personnel qualifications.
3 Methods of decision support using the apparatus of fuzzy
logic
3.1 Basic tools for creating mathematical models of CIS
As a tool for creating mathematical models (MM), the method of group accounting of
arguments [7] can be used as one of the most effective methods of structural and pa-
rametric identification of complex systems based on observational data under condi-
tions of incomplete information. In order to increase the information content of the
initial data and the accuracy of solving identification and optimization problems, it is
necessary to classify objects. For classification, it is advisable to use clustering meth-
ods that allow the division of the studied set of systems into separate classes, within
which the systems are considered to be of the same type. The k-means method or
mutual absorption method is used as a clustering algorithm. Using this method of
splitting into classes allows you to obtain subsets containing objects that are homoge-
neous in terms of the values of the properties that characterize them. The selected
subsets have this property because the mutual distance (absorption radius) between
any pair of objects in the subset is less than the distance between two objects in dif-
ferent subsets.
Then the search for optimal alternatives in its class is carried out using the Pareto
method. The choice of the optimal from this set of alternative options is possible us-
ing an additional criterion. For this, a compositional method for choosing alternatives
is being developed. It combines the solution of the problem of optimizing the objec-
tive function and the reduction of private preferences to general requirements, which
makes the choice more objective and accelerates the calculation.
The objective function in the construction of a corporate information system is an
efficiency criterion, which will be built using the developed method of forming an
integral efficiency criterion using a set of models with signs of their significance, thus
increasing the degree of certainty in decision-making.
The decision-maker, in the course of his activities, proceeds from experience, intui-
tion and knowledge. However, these foundations are difficult to algorithmize, as a
result of which active algorithmic support of choice is required when making deci-
sions by building adaptive models.
Experts play an important role in decision-making processes. At the same time, the
competence of experts is usually unknown, and expert assessments may be biased.
Decision support technologies implemented with the use of classification technologies
and optimization of variants, as well as with the use of fuzzy logic, are effective.
3.2 Application of fuzzy logic to support decision-making
Continuous improvement of business processes leads to an inevitable change in the
conditions for the functioning of the information space of a financial organization,
which predetermines the need to adapt the corporate information system to changing
requirements, as well as to assess the effectiveness of the current CIS within its life
cycle.
The restrictions imposed by the requirements act as a criterion for assessing the ef-
fectiveness of the corporate information system. Often, such an assessment is carried
out formally, and the failure to comply with non-critical restrictions that do not reduce
the value of the corporate information system may lead to the fact that the corporate
information system will be rejected, although de facto it is still capable of performing
the specified functions within its life cycle. Therefore, in order to minimize the num-
ber of such situations, an approach to formalizing requirements using the apparatus of
fuzzy logic is proposed.
Let us denote the set of requirements for the characteristics of the n-th component
of the CIS (for example, a module that generally performs one function) as
Rnl , l 1, Ln . Usually, in practice, the requirements are formalized by specifying are
as ynl for the corresponding characteristics in the form of one-sided or two-sided
inequalities, for example, ynl L ynl ynl H . The predefined possible values can be
presented in tabular form.
If we take for r a variable that shows whether the requirement R is fulfilled or not,
for example, for r = 1, the requirement R is fulfilled, and for r = 0, it is not, then we
get a full set of alternatives for which the requirements for the CIS are fulfilled. Given
the rigidity of such a selection, we can with a high degree of probability get a situa-
tion when either there are no results at all, or they do not satisfy possible modifica-
tions or changes in the configuration of the CIS. In addition, this approach imposes
rather strict restrictions when formulating the corresponding requirements.
In this regard, it is proposed to apply fuzzy logic approaches to solving this prob-
lem, which provide a more flexible formalization of requirements. In this case, the
logical values 1 or 0 are replaced by a wider range of options given by the member-
ship function (MF) to a fuzzy set r R ( y ) . In this case, the degree of fulfillment
of the requirement is determined as the value of the corresponding membership func-
tion r [0;1] .
The quantitative characteristics of the CIS can be formalized [5]: a linear s-shaped
membership function for restrictions on the right, a linear z-shaped membership func-
tion for restrictions on the left, and a trapezoidal membership function for two-sided
restrictions.
Thus, we move from rigid binary constraints to a more flexible formalization of re-
quirements [4].
To take into account the peculiarities of meeting the requirements, membership
functions (Figure 1) to a fuzzy set of a special type are proposed:
0 ; y yB
с ; y yB
S ( y ) S ( y ; y B , yT , с ) (1 c ) y cyT y B (1)
; y B y yT
yT y B
1 ; y yT
─ Unilateral restrictions from below;
1 ; y yB
(с 1)( y y B )
; y B y yT
Z ( y ) Z ( y; y B , yT , с ) yT y B (2)
0 ;yy
с ; y yT
─ Unilateral restrictions from above.
Parameters of MF (1) and (2) are the boundary required values determined by the
initial inequality constraints, an additional value, starting from which the MF takes a
unit value and an estimate of the degree of fulfillment of the requirement when the
boundary required value is reached. The “step” in the MF (Figure 1) provides a transi-
tion through a point corresponding to the boundary required value [10].
MF serves as a transition from strict to softer restrictions, allowing the decision
maker to expand the choice of alternatives within the framework of improving the
CIS. It becomes possible to compare various characteristics and more flexible deter-
mination of their compliance with the established requirements [6].
Taking into account the need to fulfill all the requirements and the impossibility of
compensating for non-fulfillment of some requirements by fulfilling others, let us
determine the coefficients of compliance with the requirements using the t-norm of
the "minimum" type. Let all the L requirements be ordered and the requirements
from the first to LF are functional, and the subsequent requirements from LF 1 to
L are non-functional. Then expressions for the coefficients of compliance with func-
tional and non-functional requirements take the form:
a) b)
Fig. 1. Graph of membership function: s-type (a) and z-type (b).
KF min rl K NF min rl
l 1, LF , l LF 1, L . (3)
Here rl R ( y ) is the degree of fulfillment of the lth requirement. If it is neces-
l l
sary to take into account the significance of individual requirements to determine the
coefficients of compliance of alternatives with the requirements, a weighted t-norm
[7] of the "minimum" type can be used:
K F min {1 wl (1 rl )} K NF min {1 wl (1 rl )}
l 1, L F l LF 1, L (4)
,
Where wl is the weight that determines the significance of the lth requirement.
4 Conclusion
Requirements for CIS, as well as for IS in general in a financial organization, are
constantly changing, which imposes additional requirements for its flexibility, secu-
rity and speed of adaptation. Constant changes in legislation introduced by the regula-
tor are associated with a risk that must be taken into account in the activities of the
organization. These changes can have a serious impact on business processes, and,
therefore, on the company itself.
Thus, the conceptual model of managing the life cycle of the corporate information
system, as the most subject to changes in the element of the financial organization's
IS, is the basis for creating a universal technology for building the corporate informa-
tion system. These mechanisms will make it possible to track seemingly invisible
negative trends in the degradation of corporate information systems, including within
the framework of its integration with other systems, both in monolithic and microser-
vice architectures. The ultimate goal of implementing these models is the ability to
automate monitoring of the corporate information system compliance with changing
needs, incl. in the future, and inform decision makers about the need to intervene in
the life cycle of the IS. Continuous improvement of IS will ensure its harmonious
evolutionary development within the framework of fintech business functions, and not
in isolation from them.
References
1. Kvasov, A.S.: Analysis of life cycle management processes of information systems. Mate-
rials of the XXV International Scientific and Technical Conference, 298-301 (2019).
2. Kvasov, A.S.: Methodological aspects of ensuring the compliance of the information sys-
tem with the requirements of the business customer. Nizhny Novgorod, 51-54 (2017).
3. Kvasov, A.S.: Multi-criteria analysis of the compliance of the information system with the
requirements of the business customer. Materials of the XVII International Youth Scien-
tific and Technical Conference “The Future of Technical Science”. N. Novgorod: NSTU
named after R.E. Alekseeva, 35-36 (2018).
4. Yegorov, Y.S., Milov, V.R., Kvasov, A.S., Sorokoumova, S.N., Suvorova, O.V.: Formal-
ization of software requirements for information systems using fuzzy logic. IOP Conf. Se-
ries: Journal of Physics: Conf. Series, 1015, 042062 (2018).
5. Baranov, V.G., Sedakov, D.V., Milov, V.R. Information-measuring and control systems, 8,
26–33 (2017).
6. Hull, E., Jackson, K., Dick, J.: Requirements Engineering. Publisher: Springer-Verlag
London, 227 (2011).
7. Ivakhnenko, A.G., Yurachkovsky, Yu.P.: Modeling of complex systems based on experi-
mental data. Moscow: Radio and communication, 120 (1987).
8. Dmitrieva, N.G.: Methodological models of information support management. Proceed-
ings of NSTU im. R.E. Alekseeva, 1 (116), 11-22 (2017).
9. Baranov, V.G., Sedakov, D.V., Milov, V.R.: A method for managing the configuration of
a complex of information systems. Information-measuring and control systems, 8, 26-33
(2017).
10. Skripnik, I.V., Egorov, Yu.S.: The choice of system and technical solutions in the con-
struction of radio communication equipment using the apparatus of fuzzy logic. Design
and technology of electronic means, 3, 21-27 (2019).