Topic: Method of structuring goals. Method of structuring goals: essence, goals, objectives and objects of application
Structuring goals and setting targets.
Design organizational system
Structuring goals and setting targets.
After setting goals and developing a strategy, it is necessary to structure their achievement. In other words, design a “tree” of goals - a model of subgoals (tasks), the solution of which will technologically ensure their achievement.
Design is carried out through analysis of key factors for achieving results based on:
Theories (knowledge) by type of activity;
National and foreign experience;
Developer competencies;
Information about the external environment and opportunities;
Other reasons, depending on the specifics of the activity.
Various methods and principles can be used in developing a “tree” of goals, for example:
Logical methods;
Marketing methods;
MECE principles (Mutually Exclusive, Collectively, Exhaustive) - (mutually exclusive, jointly exhaustive);
Other methods at the developer's discretion.
Objects of goals and subgoals: consumers (requests) + non-consumers, goods (supply), market (demand, competition, alternatives).
This book is about how to get out of the ocean of competition and create a new market niche for yourself that you have never even thought about before. The blue ocean strategy allows you to abandon the struggle for ever-decreasing demand in developed markets and create growing demand in new ones.
Former accordionist, acrobat and fire eater Guy Laliberte is now the head of Cirque du Soleil, one of Canada's largest cultural exporters. Created in 1984 by a group of street actors, the company has achieved incredible success. This fast growth It is also notable because it occurred not in an attractive industry, but in a declining one, where traditional strategic analysis pointed to limited growth opportunities. The power of suppliers in the form of star performers was as strong as the power of consumers. Alternative views entertainment - from a variety of city shows and sporting events to home entertainment - increasingly took the circus industry into the shadows. Children begged their parents for money for game consoles, and not for a ticket to a traveling circus performance. In part, as a result of this, the circus industry was constantly losing customers and, as a result, revenue and income were falling. In addition, animal rights groups increasingly opposed the use of animals in circus acts, Kingling Bros., and Barnum & Bailey set the tone, and competing small circuses imitated them with their own inferior versions. In general, from the point of view of competitive strategy, the circus industry looked unattractive. Another attractive aspect of Cirque du Soleil's success was that it did not win by luring away customers from the fading circus industry, which historically catered to children. Cirque du Soleil did not compete with Ringiing Bros, and Barnum & Bailey. Instead, the company created a new, unoccupied part of the market, free from competitors. He was aimed at absolutely new group consumers: adults and corporate clients who were willing to pay several times more than a ticket to a regular circus cost in order to see a new, unparalleled performance. The name of one of Cirque du Soleil's first projects said it all: “We are reinventing the circus.”
In addition, the targets are:
- Organization and management of activities;
- Production of goods, services, trading activity and other activities to create income carriers:
- Sales, sale of goods.
1.2 Business process system.“Success is a few correct actions repeated daily, not success is a set of wrong actions repeated daily" (John Ron).
To carry out activities, you need to understand: what, who is doing, with what, from what, when, where, how.
A business process is a system of consistent, targeted and regulated decisions, actions, interactions (labor), in which, through control actions and with the help of resources, process inputs are converted into process results that are valuable to consumers.
Business processes are technologies for solving problems to achieve goals and certain results.
The role of business processes in efficiency and effectiveness:
1. Combination of labor and use of resources;
2. The rationality of labor and the use of resources in business processes determines the costs of products and activities in general;
3. Advantages in processes affect the conquest of the market, consumers and product price.
A business process system is a synchronization, a synthesis of the components of business processes, which determines the economic potential of the company.
Key points in organizing business processes:
Separation and structuring of goals into the necessary areas.
Identification of performance and efficiency factors based on criteria;
Engineering (design) of business processes: operations, technology, resources, labor, estimated indicators;
Determining the rules and procedures for applying organizational and management systems;
Synthesis of business processes into a system – obtaining synergy through interaction of processes.
Synergy – (Greek – acting together) – an increase in the efficiency of activity as a result of connection, integration, merging of individual parts into a single system.
Business process design (engineering) is carried out by creating a technological model. Engineering (English - ingenuity; invention; knowledge. Engineering is a set of techniques and methods that are used to design activities in accordance with goals.
Business process engineering includes:
1. Analysis and division of activities into business processes. Determination of results (products of business processes);
2. Development of process technology for obtaining products: components of the operation, order and rules for their implementation;
3. Establishing the necessary resources to carry out the processes;
4. Determination of the content of labor in processes: decisions, actions, interaction, complexity, labor intensity;
5. Optimizing effectiveness and efficiency – productivity and costs;
6. Determination of interaction within the process and with other processes;
7. Factors and performance indicators of the process system.
The results of engineering are recorded in the organization’s documents (for example, in technological maps).
In the course of activities, business process reengineering is carried out in cases of changes in activities, strategies, tactics, introduction of innovations, decline in performance, etc.
Business process engineering technology:
Stage 1– Identification of key business processes:
a) Organization and management processes;
b) Marketing - study external environment: knowledge, information;
c) Development and improvement of goods for consumers, taking into account market guidelines, processes, labor. Introduction of innovations.
d) Production of goods, provision of services;
e) Sales – selling goods to consumers on the market.
These processes are key, as they are carried out by all forms of organizations: commercial, government, non-profit. Only the sources of payment differ.
Business process diagram.
The functional composition of key business processes is standard for the organization, and the content of business processes is unique to a particular organization.
Content is designed for each business process. The content of business processes is determined based on the structure, criteria, properties and factors.
Business process structure:
Purpose – goals (result parameters), tasks, which to achieve, why and why.
What - materials, means of production and other necessary resources.
Where – geography, land, premises.
When – timing, time.
How – technologies, methods, methods, labor.
Some - themselves, someone else.
Criteria for identifying business processes:
1. Product is the result of execution.
2. Direct or indirect participation in income.
3. Technological necessity, processes cannot be abandoned.
Properties of business processes:
Cause-and-effect relationships.
Alternativeity in execution, methods, means of execution.
Productivity (quantity).
Quality.
Complexity.
Labor intensity.
Repeatability.
Consistency with other processes. Interaction.
Cost of the process.
Business process design includes:
1. Determination of the product (output) of a business process based on the business strategy and local strategies.
2. Development of technologies for obtaining the product.
Stage 2 – Development of action technologies for business processes.
Technology (Greek: Art, skill).
Technology development is carried out taking into account the component operations (performed sequentially and in parallel).
Technologies can use both their own developments and borrowed ones.
Development of methods includes:
1. Selection of techniques and methods of work. Determination of necessary working conditions.
2. Selecting the degree of automation.
3. Establishing the level of standardization and certification.
4. Monitoring innovations and their implementation in the process.
Approved technologies are formalized in the organization's documents.
The essence of the structuring method is to build a “decision tree” through a multi-stage expert survey using the so-called Delphi procedure. As a result of each stage of the expert survey, elements of one level of the “decision tree” are formed, their comparative assessment is made, and elements that do not correspond to the accepted selection indicators are discarded. The remaining elements undergo further structuring at a lower level. This approach makes it possible to most fully take into account and evaluate all the experts’ proposals. Each time, experts are offered elements of only one level of the “tree” for evaluation, which reduces the dimension of the problem and increases the validity of the experts’ judgments. With this approach, the process of constructing a “decision tree” is accelerated, since elements of each level are evaluated according to accepted selection indicators, and elements that do not correspond to them are discarded and not taken into account in further consideration.
Let's consider the application of the method of structuring decision making using the example of determining the scope of work on new problems of a scientific and production organization (NPO).
Required:
¨ evaluate possible ways to solve the problem, i.e. determine which specific research and design work and organizational arrangements should be carried out;
¨ assess the resources required for this, i.e. time and cost of solving the problem.
To compare options for solving a scientific and technical problem, for example, two groups of quantitative and qualitative indicators given in table can be used. 7.11, 7.12.
Table 7.11.
Indicators for evaluating the elements of the "decision tree"
At the sub-problem level
| Index | Possible values | Score (points) | ||
| Compliance with the scientific profile of the NGO h (1) | Compliant Does not comply | |||
| Focus h (2) | Solves the main task of the problem Contributes to the solution of the problem along with other sub-problems Contributes little to the solution of the problem | |||
| Novelty h (3) | A subproblem represents the development of a new object of technology or scientific research. A subproblem represents the improvement of an existing object of technology or scientific research. | |||
| Openability h (4) | The results of solving the problem can be implemented: on an industry scale at several enterprises in the industry in a specific research and development organization | |||
| Prospects h (5) | The results of solving the sub-problem will serve as the basis for new research work. The results are important only for solving this problem. | |||
Table 7.12.
Indicators for assessing the elements of the “decision tree” at the research level
The procedure for structuring the problem and its assessment involves the implementation of the following stages:
stage ] - definition and description of the problem;
stage 2 - structuring the problem at the subproblem level;
stage 3- structuring sub-problems at the topic level scientific research;
stage 4- calculation of technical and economic characteristics of elements of the third level of the “decision tree”.
The definition and description of the problem are discussed above, so let’s move on to the second stage.
Stage 2. Structuring the problem
at the sub-problem level
Step 1. A group of experts is formed so that it includes representatives of all main areas of work in this NGO. Experts become familiar with the problem and formulate areas of research and development (sub-problems) that can lead to a solution to the problem at hand.
Step 2. The expert competence coefficient is calculated.
Each expert indicates the degree of his awareness of the problem on a ten-point scale and the sources of argumentation for his opinion. The expert’s reasoning is determined by summing up the points according to the reference table. 7.1.
The expert receives such a table without numbers and notes (with a sign) the degree of influence of each source on his opinion. After applying the reference table, the sum of points is calculated for all sources noted by experts.
Step 3. Compiled common list sub-problems presented by all experts (duplicates are excluded). A generalized expert assessment of the compliance indicator is calculated jth problem NGO profile:
If h j - (1)< ½, то jth problem excluded from further consideration.
Step 4. Obtaining from each expert an assessment of the subproblem feasibility indicator h (2) can take values 100, 50, 10 (Table 7.12). To establish the degree of agreement between expert opinions on the appropriateness of the identified problems, the concordance coefficient is calculated. To do this, experts rank subproblems in such a way that the subproblem that most closely matches the solution to the problem receives a rank of 1, the next one receives a rank of 2, etc. Subproblems that are equivalent, according to experts, receive a rank equal to the arithmetic mean. The degree of agreement between expert opinions is characterized by the concordance coefficient w, which is determined by the formulas:
where s is the sum of squared deviations of the sums of ranks obtained j-m direction of research, from the arithmetic average of the sums of ranks obtained by all areas of research;
P - number of experts;
T- number of directions;
r j - rank assigned i-th expert jth direction;
r̅ - estimation of mathematical expectation.
If there are related ranks, the concordance coefficient (w) is calculated as follows:
Where T i - indicator of related ranks in the i-th ranking;
N i- the number of groups of equal ranks in the i-th ranking;
h K - number of equal ranks in K-th. group of related ranks when ranking by the i-th expert.
The concordance coefficient w can take values 0 £ w£1.
If w< 0.5, then a second meeting of experts is held in order to achieve the same understanding among them of the essence of the subproblems, while they can change their view of the subproblem, and when recalculating the value w may increase. If w> 0.5, then move on to the next step.
Step 5. Calculation of a generalized expert assessment of the focus indicator j-and subproblem h j - (2) :
where h j - (2) is the assessment of the i-th expert of the indicator feasibility j-th subproblem, which can take the value 100, 50, 10 (Table 7.11).
If h j - (2) < 30, i.e. Most experts believe that the jth subproblem contributes little to solving the problem, then it is excluded from further consideration.
Step 6. Experts arrange the values of the coefficients of novelty (h (3)), implementability (h (4)), and prospects (h (5)) in descending order of their importance to determine whether the experts’ approach to assessing indicators is the same. In this case, the concordance coefficient is calculated to assess this problem, then weighting coefficients in fractions of unity are assigned to each of these indicators
Step 7 Calculation of the average value of weighting coefficients for indicators of novelty, implementation, prospects ():
Where l- indicator number ( l = 3, 4, 5);
q ij- weighting coefficient of indicator h (1) given by the i-th expert;
P- number of experts.
Step 8 Each expert's assessment of the subproblem according to indicators h (3), h (4), h (5), while h (3), h (4) can take values 100, 50, 10, and h (5) takes values 100, 50 (see Table 7.11). The results of the survey of experts on the indicators h (3), h 4), h (5) can be conveniently presented in the form of a table. 7.13, where x ijl - the value of the indicator h (1) assigned to the j-th subproblem by the i-th expert.
Step 9 Calculation of the significance level indicator (х̅ j), i.e. generalized meaning jth grades subproblems according to indicators h (3), h (4), h (5). At this step, the group opinion of experts about jth value subproblems in solving the problem as a whole:
Table 7.13.
Expert ratings for indicators h (1)
where x̅ ̅ j-
q̅ ̅ j- average value of weighting coefficients for each indicator;
x ijl - the value of the indicator h (1) assigned to the j-th problem by the ith expert (100, 50, 10).
Step10. Order in descending order of significance level (x̅ jmax = 100, x̅ jmin = 10) and calculation of the importance coefficient for each subproblem K bj(in fractions of a unit). The importance factor shows the relative jth importance problem to solve the original problem, when calculating it, the level jth significance sub-problems and the degree of its focus:
where x̅ ̅ j -- level of significance of the j-th subproblem;
h j - (2) - the average value of the expert assessment of the focus of the j-th subproblem.
Stage 3. Structuring sub-problems at the level
research topics
At this stage of calculations, the elements of the third level of the “decision tree” are determined, i.e. directly topics of scientific research. In this case, it is possible to change the composition of the expert group. It includes specialists primarily from those fields of activity that correspond to the problems selected for further consideration.
Step 1. Formation by experts for each j-th subproblem of a list of proposed topics of scientific research and expected results. An expert can indicate research topics on one or more sub-problems and evaluate K i, K a(coefficients of awareness and argumentation), as is done in step 1 of the previous stage. This data is necessary to determine the expert’s competence coefficient (see step 2 of the previous stage). In addition, the expert must indicate the proposed executor (division, responsible executor) for the proposed research topics.
Step 2. Analysis of the list of proposed topics to identify those adequate to the expected results and the composition of performers to identify those adequate to the research. In the case of topics that are identical in content, the wording is clarified and one of them is included in the general list of topics.
Step 3. Evaluation of research topics according to the focus indicator (h (2)) Kth topics when solution j-th Problems; h (2) can take values 100, 50, 10 (Table 7.11). The results of a survey of experts on indicator h (2> can be conveniently presented in the form of a table, where x ijk - value of indicator h<2) , присваиваемое Kth topic by the i-th expert on the j-th subproblem (Table 7.14).
Step 4. Determination of a group expert assessment of the focus of the K-th topic when solving the j-th subproblem (A̅ jk):
Where X(2)ijk - focus assessment Kth topics for solving the j-th subproblem, given by the i-th expert;
K i-- coefficient of competence of the i-th expert, normalized to unity;
i - expert number;
j - subproblem number;
TO - topic number (TO= 1, 2, ..., L);
A̅ jKmax= 100, A̅ jKmin = 10.
The method of structuring goals provides for a quantitative and qualitative description, deadlines for achievement and analysis of hierarchically distributed interrelated and interdependent goals of the management system. Structured goals are often presented graphically as a “tree” of goals, showing the connections between them and the means to achieve them. The construction of such a “tree” is carried out on the basis of deductive logic using heuristic procedures. It consists of goals of several levels: general goal - main goals (subgoals of the 1st level) - goals of the 2nd level - subgoals of the 3rd level and so on until the required level. To achieve the general goal, it is necessary to realize the main goals (essentially, these goals act as a means in relation to a higher goal); to achieve each of the main goals, it is necessary to implement, respectively, its more specific goals of the 2nd level, etc. Typically, classification, decomposition and ranking procedures are used to build a “tree” of goals. Each subgoal should be characterized by a coefficient of relative importance. The sum of these coefficients for the subgoals of one goal should be equal to one. Each level of goals (subgoals) should be formed according to a certain criterion for decomposing the process of achieving them, and any goal (subgoal) should preferably be attributed to an organizationally separate unit or executor.
Normative method
The normative method involves the use of a set of certain established standards, comparison with which real indicators of the control system can make it possible to establish the compliance of the system, for example, with the accepted conceptual model. Standards can determine the composition and content of functions, the labor intensity of their implementation, the number of personnel, the type of operating system, etc. An example of standards can be the controllability standard, the number of permissible management levels, labor intensity standards for performing functions, etc.
Parametric method
The parametric method is based on the quantitative expression of the studied properties of the control system and the establishment of relationships between the parameters of the control and controlled subsystems. This makes it possible, on the basis of actual data, to determine the form of dependencies of interrelated parameters and their quantitative expression. The dependencies of the parameters can be functional (manifested definitely and accurately in each individually observed case - observation) or correlational (determined on the basis of the correlation method).
Correlation (interrelated) method
The correlation (interrelated) method is one of the economic and mathematical research methods that makes it possible to determine the quantitative relationship between several parameters of the system under study. In this case, the correlation dependence, unlike the functional one, can manifest itself only in the general, average case, that is, in the mass of cases - observations.
The initial most important task of the correlation method is to determine the type of correlation equation (regression equation). The simplest form of such an equation, characterizing the relationship between two parameters, can be the equation of a straight line
Y = a + in X,
where X, Y are the independent and dependent variables, respectively;
a, b are constant coefficients.
The conclusion about the rectilinear nature of the dependence can be verified by simply comparing the available data or graphically (by registering in a rectangular coordinate system the values of Y and X, the location of which on the graph allows one to draw a conclusion about the correctness or incorrectness of the idea of the linear nature of the dependence between the two parameters being studied).
The next task is to determine the constant coefficients of connection between variable parameters that will best correspond to the available actual data Y and X. In this case, as a criterion for assessing the adequacy of the linear dependence to the actual data, you can use the minimum sum of squared deviations of the real statistical values of Y from those calculated using the accepted equation for direct use. The coefficients of the line when using this criterion can be determined by the well-known least squares method.
An example of a linear relationship is the number of deputy heads of a functional department on the number of employees X in the department and, based on statistical data (for this example, usually at least 20-25 pairs), obtain the following relationship
Y = O.600 + 0.206 X.
The value of the parameter under study is quite often influenced by not one, but several factors. In this case, for example, with a linear connection between all factors, you can use a linear multiple correlation equation of the following form
Y = a 0+a1X1+a 2X 2 +...+a nX n.
If the impact of any factor on the object under study cannot be considered linear, then the corresponding factors can be included in the equation not in the first, but in a higher degree, for example, in the second:
Y = a 0+a1X1+a 2X2+a 3X 23.
Correlation methods, especially with multiple correlation of parameters, are effectively used, as a rule, in conjunction with other economic and mathematical methods.
New economic relations and a qualitatively different management system require the use of more modern methods of analysis and system design. One of these methods should include the FSA method, which, as research shows, can be adequate to modern requirements when carrying out research work on improvement.
In any case, regardless of the specific method used, the greatest effect and objectivity of research work can be achieved by the comprehensive use of methods acceptable for the purposes of CS research. Moreover, some of them may be effective at one stage of the study, while others - at another.
Any organizational activity in the system is justified only if it contributes to achieving its ultimate goal of functioning. In other words, any organization must be designed in such a way that all organizational activities in the system realize only those operational goals for which it was created. At each level of the organization, private goals arise, and only their totality must be considered as a certain goal of a certain level of management. Hence the need to structurize goals and build a tree of goals.
Goals- This is a specification of the organization’s mission in a form accessible when managing the processes of their implementation. They are characterized by the following features and properties:
Clear orientation to a specific time interval;
Specific and measurable;
Consistency and consistency with other goals and resources;
Targeted and controllable.
Designing the organization's goals is a mandatory and important step in the overall management system design process.
Stage 1. Researched the purpose of the organization, the overall goal of the system, determined by its purpose. When analyzing, it is important not to confuse the concepts of the intended purpose of the system and the criterion of effectiveness. In contrast to the purpose of the organization, determining the direction and meaning of the functioning of the system, its main task, the effectiveness criterion is an indicator that determines the degree to which this goal is achieved.
Stage 2. Are being formed quality goals of the organization, carried out in two stages. First, qualitative goals for the functioning of the organization are formed, arising from the Regulations of this organization that regulate its activities, and then the qualitative characteristics of the goal are selected. The formulation of the goal should be extremely clear and indicated in the imperative mood.
Stage 3. Produced ranking the goals of the organization's functioning, carried out to justify the choice of operating goals that form the basis for further design of the organization. Since all real organizations are multi-level, the ranking of goals should be carried out at all selected levels of management. Ranked goals don't just happen. They arise from global operational goals, which, in turn, are regulated by decisions of higher organizations, regulations, and regulations on the organization.
On fourth stage carried out building a goal tree. The number and variety of management objectives are so great that no organization, regardless of its size, specialization, type, or form of ownership, can do without an integrated, systematic approach to determining their composition.
30. Study of parameters of organizational management structures
The organizational structure of management is an integral set of elements of an object and a management body interconnected by information links. It reflects the structure of the management system, the content of which is management functions, the vertical and horizontal relationship between management levels, as well as the number and relationship of structural units within each level. Depending on the relationship between levels and structural units, linear, functional, linear-functional, matrix and matrix-staff types of organizational structures are distinguished.
Currently, classical linear-functional structures are typical only for small and some medium-sized companies. They are rarely used at the level of transnational corporations, more often at the level of their divisions abroad. For large companies, the divisional approach to building organizational management structures has become dominant.
Divisional (departmental) structures management (from the English word division - department, division of a company) are the most advanced type of organizational structures of a hierarchical type and are even sometimes considered a cross between bureaucratic (mechanistic) and adaptive structures. In some cases, these structures can be found in the literature under the name “fractional structures”.
Divisional structures - structures based on the allocation of large autonomous production and economic units (departments, divisions) and the corresponding levels of management, granting these units operational and production independence and transferring responsibility for making a profit to this level.
While creating consumer-oriented organizational structures, divisions are grouped around certain groups of consumers (for example, the army and civilian industries, products for industrial, technical and cultural purposes). The goal of such an organizational structure is to satisfy the needs of specific customers as well as a company that serves just one group of them does. An example of an organization that uses consumer-oriented management structures is commercial banks. The main groups of service consumers in this case will be: individual clients, companies, other banks, international financial organizations.
Project structures- these are management structures for complex activities, which, due to their decisive importance for the company, require the provision of continuous coordinating and integrating influence under strict restrictions on costs, timing and quality of work.
Traditionally, a department manager in any large company within a hierarchical organizational structure has many different responsibilities and is responsible for various aspects of several different programs, issues, projects, products and services. It is inevitable that under these conditions, even a good leader will pay more attention to some types of activities and less to others. As a result, the inability to take into account all the features and all the details of projects can lead to the most serious consequences. Therefore, in order to manage projects and, above all, large-scale ones, special project management structures are used.
One of the most complex management structures of the adaptive type is recognized matrix structure. It was originally developed in the space industry and has been used in the electronics and high-tech industries. The matrix structure arose as a response to the need for rapid technological change while making the most efficient use of a highly skilled workforce.
Matrix structure reflects the consolidation in the organizational structure of the company of two directions of leadership, two organizational alternatives. Vertical direction - management of functional and linear structural divisions of the company. Horizontal - management of individual projects, programs, products, for the implementation of which human and other resources of various divisions of the company are involved.
1 Method of structuring
1.1. Hierarchical structures and goal tree
The idea of the goal tree method was first proposed by W. Cherman in connection with the problems of decision making in industry.
The term “tree” implies the use of a hierarchical structure (hence the name “structuring method”), obtained by dividing the overall goal into subgoals, and these, in turn, into more detailed components, which can be called subgoals of lower levels or, starting from a certain level , - functions. As a rule, the term “goal tree” is used for hierarchical structures that have strictly tree-like relationships, but the method itself is sometimes used in the case of “weak” hierarchies. Therefore, recently the term “forecast graph” proposed by V.M. Glushkov, which can be represented both in the form of a tree-like hierarchical structure and in the form of a structure with “weak” connections, has become increasingly widespread.
When using the goal tree method as a decision-making tool, the term “decision tree” is often introduced. When using a “tree” to identify and clarify management functions, they speak of a “tree of goals and functions.” When structuring the topics of a research organization, it is more convenient to use the term “problem tree”, and when developing forecasts, the term “tree of development direction (or development forecasting)” or the above-mentioned term “forecast graph”.
The “goal tree” method is aimed at obtaining a complete and relatively stable structure of goals, problems of directions, that is, a structure that has changed little over a period of time with the inevitable changes occurring in any developing system. To achieve this, when constructing structure options, one should take into account the patterns of goal formation and use the principles and methods of forming hierarchical structures of goals and functions.
To successfully apply this method, three main types of input data are required:
1. clearly defined goals, objectives, systems and their components at all levels;
2. interrelated criteria for measuring the relative importance of components at each level;
3. numerical assessments of significance according to the criteria of each level.
It should be noted that the relationship of tasks in the goal tree is established regardless of the probability of intermediate outcomes and possible solutions; it does not take into account that the exclusion or addition of several intermediate links has an impact on the work program as a whole.
Another serious difficulty is associated with the need for a numerical assessment and synthesis of various technical, time and cost characteristics of alternatives, which is poorly achieved when using the goal tree principle.
To eliminate some of these selection difficulties, the branching tree principle, which is process-oriented rather than goal-oriented, can be used. Process orientation provides an analysis of the dynamics of successive stages of the program, taking into account the probabilistic outcomes of each stage.
However, in practical activities, a significant part of the work is qualitatively new and insufficiently defined with regard to the technical implementation of costs and deadlines. In all cases, a complex logical situation arises when each job is a random variable, and the occurrence of each of the expected network events depends on the probability of the occurrence of previous events and on external conditions.
The analysis of such situations can be performed using decision trees that provide modeling of complex situations that arise when choosing areas of scientific research, development options and capital investments. The decision tree includes options for action, as well as possible events and results of actions that are influenced by chance and factors beyond our control. Naturally, the results of various decision options are based on the information available to us at the time of decision making. Despite the fact that some of these events will not materialize, when making a decision on the choice, it is necessary to assess the likelihood of their occurrence.
Such estimates can be summed, allowing the conditional probability of achieving each possible outcome to be calculated. When analyzing problems, these results can be expressed in terms of the expected cost of each action or possible results of the work.
In addition, with the help of such a tree, in a complex chain of decisions, one can take into account the time and cost factors by analyzing the tree, starting from the last decision in the direction opposite to the flow of time, up to the initial decision and assessing the relative importance of each node of the tree as the difference between the expected costs for its achievement and expected results.
1.2. Decision tree structure
Tree branches are arcs (activities) of a network with two or more end nodes (events). Nodes are states in which the possibility of choice arises, both due to the actions of the decision maker and due to the influence of external, uncontrollable factors (“nature”). In decision tree diagrams, squares indicate nodes where the choice is made by the decision maker, and circles indicate nodes where the choice depends on the influence of external conditions.
The sequence of the procedure for selecting the most preferable alternatives using a decision tree can be presented in the form of the following main stages:
1. analysis of the problem, that is, the establishment of possible solutions that can be made and factors that may influence the results of decisions;
2. assessing the probability of each of the network events and calculating the total probability of each outcome;
3. distribution of costs by type of work and assessment of the cost of “delay”;
4. consistent reassessment of events taking into account preliminary results.
Approximate structure of a decision tree:
holding P1=0.4
research
Problem
p6=0.6 does not occur
the problem does not arise| |
1,2,3,4,5 - possible outcomes
1 – problem resolved, minor delays and cost overruns,
2 – problem solved, additional costs and delays,
3.5 – project completed,
4 – problem solved, large cost overruns and delays, project completed.
Having established the probabilities of the stages along each path to solving the problem, we can calculate the conditional probabilities of each possible outcome occurring.
Thus, the conditional probability of outcome one will be equal to P1*P5=0.4*0.2=0.08, the probability of outcome 2 will be P1*P6=0.4*0.8=0.32, and the probability of outcome 3 will be 0 ,6. the total probability of an alternative associated with carrying out research work is 0.08+0.32+0.6=1.0. similarly, the total probability of an alternative associated with replacing one of the units with a more advanced one is also equal to one Р3+Р4=0.5+0.5=1.0
Also using the tree, it can be noted that conducting research would reduce the likelihood of delays associated with additional costs of money and time from 0.5 to 0.32.
In cases where the expected profit can be estimated, all costs and revenues (or losses) are discounted and multiplied by the probability of success of the alternative branch of the decision tree, which allows us to establish the expected “price” of the alternative solution.
