<<
>>

the Choice of programmno-analytical toolkit of an estimation of efficiency of the organisation of business processes of the enterprise

The estimation of efficiency of the organisation of business processes of the enterprise allows to reveal directions of their reengineering that will supply removal of "narrow" (problem) places on production on the basis of application of the most effective technology of modelling of business processes.

The estimation of efficiency of the organisation of business processes of the enterprise should be conducted with allowance for applications of toolkit of "through" management which includes the expanded set adequate to modern conditions of managing of the tools, allowing to evaluate not only an organisation strategic location as a whole, but also covers all stages of complete life cycle of industrial system.

Tools of "through" management concern [51]:

1. Strategic methods and models

2. Traditional methods and models

3. Toolkit of analytical engineering

4. Inzhiniringovye methods and models

5. The strategic structured set of accounts

6. A situation analysis

7. Stsenarnyj the analysis

8. prognoznyj the analysis

9. A factor analysis

10. The segmentary analysis

11. The social analysis

12. Fraktalnyj the analysis

13. The investment analysis

14. The innovative analysis

15. portfelnyj the analysis

16. The tactical analysis

17. The competitive analysis

18. The functional analysis

19. The strategic control.

The analysis of practice of application of methods of "through" management on an example of the enterprises of the Peoples Republics of China realising complete life cycle of a product (Appendix), proves an urgency and introduction of "through" management at the enterprises of complete life cycle of the Peoples Republic of China which are one of the advanced countries in the field of introduction of the advanced managements-technologies [52].

The analysis of efficiency of the organisation of business processes should be conducted also from the point of view of an estimation and forecasting of finance results of activity of the enterprise. Now wide enough toolkit is applied to forecasting of indicators of development of industrial enterprise. In forecasting of financial indexations set of special methods and receptions which can be divided into three groups is applied: methods of expert judgements, extrapolation methods, methods of economic-mathematical modelling.

The method of expert judgements is based on processing of opinions of experts concerning dynamics of the financial processes revealed by realisation of special procedures (questioning, interviewing). The qualified professionals professionally attending to studying and (or) management by economy and the finance of firm [53] should be experts.

The extrapolation method is based on distribution on the future of the tendencies which have developed in a retrospective show. Degree of applicability of a method of extrapolation in a financial field is defined by degree inertsionnosti (or stability) dynamics of development of economic system. Less inertsionny micro-economics financial indexations, therefore at level of managing subjects to a lesser degree are applicable. More inertsionna dynamics of development of financial indicators at macroeconomic level, and in these conditions applicability of a method of extrapolation increases.

The extrapolation method, as a rule, is applied to forecasting of system of financial indexations in a complex combination to other methods.

Methods of economic-mathematical modelling base on construction of models which probabilities with defined degree describe dynamics of financial indexations depending on the factors influencing financial processes. Optimistic, pessimistic and most probable rates of changes of economic indicators (gain growth, a cost-saving per unit of production, invariable tax rates, a constant share of payments in the budget) [54] are thus used.

In theory and financing activities practice the increasing significance is acquired by the methods of calculation united under the general name «the financial mathematics», or the maximum financial calculations, or financial and commercial accounts. Thus one of effective methods of reception of authentic results is business modelling - construction of the business model considering specificity of development of the organisation in modern conditions of managing.

The concept organisation "business model" is in present time one of the least unequivocal and structured terms in the modern scientific literature. In wide understanding the business model is a method of realisation of business which supplies to the enterprise the income and profit.
The business model formally reflects earning money process, in details defines its disposition and a role in a chain of creation of cost [55].

According to the researcher P.Timmersa, business model consists of set of products, services and information streams, and also the description of various participants of business process, their role in a chain of value, potential benefits with decoding of sources of acquisition of income. For understanding of business mission of the company the marketing model which unites business models and marketing strategy of the required representative of business [56 [57] is added.

A.Ostervalder, the businessman, "revolutionary" in the field of modelling business - of processes, and Willows Pine, the American scientist in the field of information technologies and systems, the professor, in teamwork «Construction of business models. The reference book of the strategist and the innovator» notice, that the business model interprets how the enterprise creates,

57 transports and realizovyvaet value.

Thus the business model is the plan of how the company strategy should be realised within the limits of its internal structures, processes and systems. The estimation of efficiency of the organisation of business processes should be continuous, built in system of "through" management, to be conducted on the basis of constant monitoring of occurring changes. For this reason forecasting of finance results of activity of industrial enterprise as the estimation of efficiency of the organisation of business processes should be conducted not only on the basis of construction effective from the point of view of reengineering of business model of activity of the organisation, but also should to update constantly in the automated mode results of the forecast with
The account of automatic updating of basic data - source parametres, that is the business model should be adaptive, trained.

The most effective business model in this case, according to the author, is the neural network. Artificial neural networks (INS) are mathematical models, and also their program or hardware realisations constructed by a principle of the organisation and functioning of biological neural networks - networks of nervous cages of a live organism. [58]

Artificial nejron is the knot of the artificial neural network which are simplified model natural nejrona. Matematicheski iskustvennyj nejron represent as a certain nonlinear function from unique argument - a linear combination from all source signals, thus received result is sent on a unique exit (fig. 10). [59 [60]

Drawing 10 - Structure artificial нейрона60

The first INS have been realised in the form of electronic schemes. After a while, INS began to be realised in the form of programs in connection with computer facilities development.

The neural network is set of the big number of rather simple elements - nejronov which topology of connections depends on network type. To create a neural network for the decision of any specific target, it is necessary to choose, how it is necessary to connect nejrony with one another, and in conformity to pick up significances of weight parametres on these communications. Influence of one element on other, depends on the established connections. Force of influence is defined by weight of connection [61].

Usually, the neural network is used when the exact kind of communications between inputs and exits as if it was known is unknown, communication could be simulated directly. Other essential feature of neural networks is dependence between an input and an exit: it is in process of network training.

Neural networks on structure of communications can be divided on:

1. polnosvjazannye - each nejron transmits a target signal to the rests nejronam and to itself(himself), all source signals move all nejronam. As target signals of a network can serve all or some target signals nejronov after several cycles of functioning of a network.

2. nepolnosvjaznye neural networks are subdivided on single-layered and multilayered, with straight lines, cross and feedback. nejrony j - ogo a layer on inputs can communicate only with nejronami underlaying layers in neural networks with direct ties.

In neural networks with cross communications connections in one layer are supposed. Among multilayered neural networks allocate:

1. Monotonous networks. Here each layer except the day off (last) is broken into two blocks: exciting and braking. Communications between blocks are divided as. If from nejronov the first block to nejronam the second conduct only exciting communications, any target signal of the block is
Monotonous not decreasing function of any target signal of the first block. If these communications only braking any target signal of the second block is not increasing function of any target signal of the first. Monotonous dependence of a target signal nejrona from parametres of source signals is necessary for nejronov monotonous networks. [62]

2. Networks without feedback. In such networks nejrony a source layer receive source signals, will transform them and transmit nejronam the first latent layer, and so on up to the day off who issues signals for the user and the interpreter. Each target signal of q th layer moves on an input of all nejronov (q+1)-ro a layer if not it is stipulated opposite, but the variant of connection of q th layer with any (q∣p) - ∖ι a layer is possible also.

3. In networks with feedback the information is transmitted on the previous layers from the subsequent. Among them allocate:

- Sloisto-cyclic - layers are closed in a ring: the last layer transmits the target signals to the initial. All layers are equal in rights and can not only receive source signals, but also issue days off;

- sloisto-polnosvjazannye consist of the layers representing polnosvjaznuju a network. Here are transmitted not only from a layer to a layer, but also in a layer. The work cycle breaks up to three parts in each layer: reception from the previous layer of signals, an exchange of them in a layer, development of a target signal and transfer to the subsequent layer;

- Polnosvjazanno-layered, on the structure are identical with sloistoyopolnosvjazannymi, but function differently: in them phases of an exchange in a layer and transfer to the following obedeniny, nejrony all layers accept signals from nejronov the and the subsequent layers on each step [63].

The behaviour of an artificial neural network depends not only on significance of weight parametres, but also from excitation function nejronov. There are three principal views of function of excitation:

1. Threshold - for threshold elements on one of two levels the exit depending on that is established, it is more or less total signal on an input nejrona than certain threshold significance.

2. Linear-target activity for linear elements is proportional to the total weighed input nejrona.

3. Sigmoidalnaja - in process of change of a source signal the exit varies continuously, but not linearly. Sigmoidalnye elements have more similarities with real nejronami, than linear or threshold, however any of these types can be considered only as approach. [64]

At training of the neural networks intended for forecasting of financial time numbers, use the standard approach. Available examples divide into three samples: training, validatsionnuju and test. Training sample is used for fine tuning sinapticheskih factors of trained neural networks for the purpose of approach to an error minimum on a network exit. Validatsionnaja sample is used at a choice of the best network from several trained, and also for definition of the moment of the termination of training. The test sample which has been not used in the course of training, serves for forecasting quality assurance. [65]

For operated training of a network it is necessary to prepare a set of training data. Such data represent examples of source data and the exits corresponding to them. The Network studies to establish connection between the first and the second. As a rule, training data undertake from historical data [66].

Further the neural network is trained by means of some algorithm of operated training at which available data are used for updating of scales and threshold significances of a network so that to minimise a forecast error on training set. In case the network is trained well, it acquires ability to model unknown function which connects significances of source and target variables, and subsequently the network is used for forecasting in a situation when target significances are unknown.

Process of training of a neural network is carried out iteratsionno, according to algorithm of return distribution of an error. At the first stage of each iteration data of the next training example act on nejrony a source layer and extend from the first layer to last, thus target significance of each nejrona is calculated under the formula 1:

Where OUTq, OUTp - target significances nejronov q and p accordingly;

fa - Activation function;

wpq - Weight factor of communication between nejronami p and q.

At the second stage of iteration of training there is a recalculation of weight factors of neural communications under the formula 2. Recalculation is made, since the last layer and finishing the first:

Where wpq (i+1) - new significance of weight factor of communication between nejronami p and q;

wpq - Old significance of weight factor of pq-communication;

The collection: the System analysis in designing and management the Collection of proceedings of XXII International scientifically-practical conference. St.-Petersburg, 2018. With. 175-187.

51

n - speed of training;

δq - Delta - factor nejrona q;

OUTp - Target significance nejrona p.

The delta-factor, participating in account of weight significances, for a target layer is calculated under the formula 3, and for the latent layers under the formula 4 [67]:

Specificity of object of research imposes certain features on use nejrosetej for the analysis of efficiency of realisation of an advertising campaign. Such feature is the choice of function of an error of the neural network which is distinct from traditional root-mean-square. The predata processing directed on increase of joint entropy of source variables, reduction of dimension of inputs of a network and normirovku source and target data is one of the important components of the analysis of data by means of neural networks [68].

Because the estimation of efficiency of the organisation of business processes represents the automated process of constant monitoring of occurring changes application effective programmnoyoanaliticheskogo a product is necessary.

The carried out research allows us to assert, that the uniform approach of a choice of programmno-analytical toolkit of business processes of the enterprise is not produced. For a choice of a programmno-analytical product it is necessary to be guided by both cost, and field performances. Procedure of a choice programmno-analytical

Toolkit of an integrated estimation of efficiency of the organisation business - of processes of the enterprise it is presented in drawing 11.

It is made by the author

Drawing 11 - the Choice of programmno-analytical toolkit of an integrated estimation of efficiency of the organisation of business processes of the enterprise

The given approach is universal and can be applied at a choice of various software products. Approbation of the offered approach us was conducted for such the most demanded now and perspective software products in the future, as Business Studio, Deductor Studio, Bizagi Modeler, BPwin, Gretl.

The system of business modelling Business Studio [69] supports a complete cycle of working out and optimisation of a control system by the company: designing - introduction - the control - the analysis, allowing to decide the such

Problems, as formalisation of strategy and the control of its achievement; designing and optimisation of business processes; designing of organizational structure and the list of staff; formation and distribution among employees of the regulating documentation; introduction of system of a quality management according to ISO standards; preparation for automation and formation of technical projects on introduction of information systems.

Deductor Studio - a job of the analyst. It is intended for visual designing of logic of decision-making. All actions are adjusted by means of only 4 masters: import, export, processing and visualisation. Studio allows the analyst to automate routine operations on data processing, to concentrate on intellectual work and formalisation of rules of decision-making.

The possibilities of software product Deductor Studio connected with forecasting of finance results of activity of the enterprise:

1. A convenient semantic layer of storehouse of data for extraction of the information with application of habitual business terms.

2. The OLAP-module: cross-countries-tables and cross-countries-diagrammes for

The multidimensional analysis of data.

3. Tens methods of the visualisation, allowing to carry out the prospecting and comparative analysis, to reveal tendencies.

4. Revealing of errors: the built in algorithms of search of admissions, anomalies, duplicates and contradictions, detection of noise.

5. Correction of errors in data on the basis of algorithms of machine training, statistics or under rigid rules.

6. Self-trained algorithms and machine training: trees of decisions, neural networks, samoorganizujushchiesja cards, associative rules.

7. The analysis of time numbers: revealings of seasonal prevalence, a trend and a casual component.

Set of methods of an estimation of quality of models with possibility of a choice of the best.

Bizagi is the BPM-system developed and directed on modelling, performance, automation and the analysis of business processes. System Bizagi includes 3 modules for high-grade adjustment of processes [70]:

Modeler — the full-function environment of modelling of processes in notation BPMN;

Studio — the environment of working out of business processes;

Engine — the environment of performance of processes which is accessible to users in any browser from any system [71].

BPwin - It is the software product developed by the company ltd. Logic Works. It is intended for support of process of creation of information systems. BPwin is developed enough simular, allowing to carry out the analysis, documentating and improvement business of processes. With its help it is possible to model actions in processes, to define their order and necessary resources. Models BPwin create the structure necessary for understanding business of processes, revealings of operating events and an order of interaction of elements of process among themselves [72].

Gretl is krossplatformennyj a software package for ekonometricheskogo the analysis which is the open, free and free software. You can extend and-or modify it according to licence GNU General Public License (GPL),
Which text is granted Free Software Foundation. Has following features [73]:

1. Set of methods otsenivanija: the least squares (LS), the maximum credibility (ML), the generalised method of the moments (GMM), a method of one equation and systems of the equations.

2. Toolkit for the analysis of time numbers: ARIMA, a wide spectrum of one-dimensional GARCH-models, VAR and VECM (including structural VAR), tests for individual roots and kointegratsiju, filter Kalmana etc.

3. Models from the limited dependent variable: logit, it is punched, models with selection displacement (Tobina, Hekmana etc.), interval regress, models schyotnyh data, duration model etc.

4. Otsenivanie panel models, including tool variables, it is punched also dynamic panels on the basis of the generalised method of the moments (GMM).

5. Issue of results in format LaTeX in the form of the table or the equation one pressing of the button.

6. Powerful built in skriptovyj language hansl, containing a considerable quantity of functions for programming and work with matrixes.

7. Cyclic structure of commands for simulations by a method Monte - Karlo and iterative procedures otsenivanija.

8. The graphic interface for thin adjustment of display of schedules gnuplot.

9. Constantly growing base of the user functional packages written on hansl.

10. Integration and data exchange with statistical packages GNU R, GNU Octave, Ox and Stata.

Gretl supports databases in following formats: own data of format XML; CSV (with various dividers); sheets Excel,
Gnumeric and Open Document; files. dta from Stata; files. sav from SPSS; working files Eviews; data JMulTi; own binary databases (support of data of the mixed rate and different length), databases RATS 4 and PC Give. At installation examples of data (on macroeconomic of the USA) are automatically granted. See also page with data sets for gretl [74].

To compare these software products it is necessary to conduct a comparative estimation of importance of criteria. Importance of criteria has been evaluated by us by paired comparisons of each factor with each other.

For five elements it is necessary to conduct comparisons, having defined, what factor and in what measure surpasses others (in a 9-mark scale of comparisons of a method of the analysis of hierarchies - MAI). We will enter designations of compared factors: А1 - reliability and safety; А2 - quality of support; А3 - funktsional; А4 - the price; А5 - convenience of the interface. In table 6 results of comparison of these factors are presented

Table 6 - the Matrix of pair comparisons of factors

Factors А1 А2 А3 А4 А5
А1 1 2 2 2 3 10
А2 1/2 1 2 3 3 9,5
А3 1/2 1/2 1 4 5 11
А4 1/2 1/3 1/4 1 1/3 2,416
А5 1/3 1/3 1/5 3 1 4,866
2,833 4,166 5,45 13 12,333 37,749

Let's calculate a vector of priorities. We will make it as follows - we will divide elements of each column of a matrix into the sum of elements of this column, i.e. we normalise a column, then we will combine elements of each received line and we will divide this sum into number of elements in line.

10/37,749 = 0,265

9,5/37,749 = 0,252

11/37,749 = 0,291

2,416/37,749 = 0,064

4,866/37,729 = 0,129

As a result we will receive a vector of priorities:

(0,265; 0,252; 0,291; 0,064; 0,129).

Let's calculate λmax. For this purpose the sum of each column we will increase by a priority and we will combine all products:

λmax = 2,833*0,252+4,166*0,252+5,45*0,291+13*0,064+12,333*0,129 = 5,44

Let's find a coordination index:

Casual index (SI) for a matrix dimension 5? 5 = 1,12

Let's calculate the coordination relation:

T to. The relation of a coordination of OS

<< | >>
A source: Lju Sjuelin. METHODICAL MAINTENANCE of "THROUGH" MANAGEMENT with DEVELOPMENT of BUSINESS PROCESSES of INDUSTRIAL SYSTEMS of COMPLETE LIFE CYCLE. The dissertation on competition of a scientific degree of a Cand.Econ.Sci. Kursk - 2018. 2018

More on topic the Choice of programmno-analytical toolkit of an estimation of efficiency of the organisation of business processes of the enterprise:

  1. THE ESTIMATION OF BUSINESS PROCESSES OF INDUSTRIAL ENTERPRISE AND THE CHOICE OF THE TOOL ENVIRONMENT FOR THEIR MODELLING AND THE ESTIMATION
  2. 2.2 Choice of the optimum tool environment of modelling of business processes of the enterprise
  3. 3.2. An estimation of economic efficiency of an estimation and a choice of organizational culture of the enterprise
  4. 3.2. An estimation of economic efficiency of an estimation and a choice organiyozatsionnoj cultures of the enterprise
  5. 2.1 Estimation of importance and problematical character of business processes of the enterprise
  6. Creation of model of business processes of the enterprise «As is» and models «As should be»
  7. 3.2. Working out and an estimation of business model of finance results of activity of industrial enterprise
  8. 3.1 Research of procedure of reengineering of business processes of the enterprise at realisation of "through" management
  9. 2.3 Technique of a choice and an estimation of efficiency of elements of a control system
  10. the Technique of a choice and an estimation of efficiency of elements of a control system
  11. 3.1. An example of an estimation and a choice of organizational culture of the enterprise
  12. 3.4 information-analytical toolkit of a technique
  13. PERFECTION OF BUSINESS PROCESSES OF INDUSTRIAL ENTERPRISE ON THE BASIS OF TECHNOLOGY OF REENGINEERING
  14. Package approach at an estimation of efficiency of advertising activity of subjects of business in a network the Internet
  15. § 1.1. The General characteristic and a modern estimation of small business in a context of efficiency national and The international legal protection
  16. characteristics of methods of an estimation of efficiency of enterprise activity used in practice in Internet advertising sphere
  17. an estimation and perfection of public management on the basis of the programmno-target approach
  18. 4.2. The Procedure of payments and a scale of an estimation of key indicators of efficiency KPI of processes of activity of the economic subject.
  19. ANDREEVA EKATERINA SERGEEVNA. Perfection of forms of investment of innovative activity of the enterprises at the organisation of operational business processes. The dissertation on competition of a scientific degree of a Cand.Econ.Sci. Irkutsk - 2014, 2014
  20. methods of an estimation of efficiency of advertising activity of subjects of business and possibility of their application in Internet advertising