Lecture
In essence, the content of a DSS considered in the narrow sense corresponds to Subsystem 3, «Specific tasks of the subject domain,» of a decision support system considered in the broad sense. The absence in it of other subsystems analogous to Subsystems 1 and 2 of the extended DSS is explained by the fact that the issues of more detailed methodological and informational support for decision-making are carried outside the DSS itself when it is considered in the narrow sense.
A distinguishing feature of a DSS is that, in the general case, the formation of decisions within it is carried out on the basis of source data describing the domain under study, as well as rules and algorithms for forming decisions that are built into the system. Nevertheless, the incompleteness of information and of the qualitative assessments of the processes under consideration, and the difficulty of formalizing certain problems, lead to the use not only of known decision-making methods but also of the additional application of knowledge and experience in the domain under study.
It is precisely for this reason that another variant of forming decisions within the system is also of interest, in which, for describing the domain under study and generating algorithms for forming decisions, knowledge and experience are additionally drawn on besides the source data and known algorithms and methods.
The implementation of this variant of forming decisions is characteristic of expert decision support systems. Expert systems are regarded as intelligent systems, which is explained as follows. If ordinary DSSs form decisions on the basis of rules built into the system by its user (the DM, a specialist, etc.), then intelligent systems, which are DSSs of a higher level, form decisions according to rules generated by the system itself, on the basis of the available knowledge base in the domain under study. An extended variant of intelligent systems is expert systems, in which decisions are formed not only through the accumulated knowledge base, but also on the basis of a dialogue between the system and its user (the DM, a specialist, etc.), who has experience in the domain under study and therefore «corrects» the algorithms for forming managerial decisions generated by the system.
Note that decisions based on the results of an expert system's work may be adopted either directly by the decision maker, or in collaboration with a specialist who conducts the dialogue with the system. Here, if the DM is a specialist in the domain under study, the expert system increases the efficiency of his work. If, however, he is not, then with the help of the expert system, through a rational distribution of functions between the human and the computer, the decision maker can achieve results that are fundamentally new for him and thereby improve the quality of his work.
Since expert systems draw on the experience of expert specialists in the domain under study, the principal methods for solving the problems considered within them are heuristic methods. At the same time, expert systems themselves model, to a greater extent, not the domain under study itself, but the mechanism of human thinking as applied to solving problems in that domain. This allows them to perform not only logical operations, but also to form reasoning and conclusions.
The expert systems described above realize the first direction of their creation — on the basis of developing special computer programs capable of generating rules for forming decisions taking into account accumulated knowledge and experience. At present, a further direction in the creation of expert systems is also developing, primarily for controlling complex technical objects and technological complexes — on the basis of complex specialized software systems that combine ready-made, intelligently interacting modules.
Comments