Solving system problems of a complex structure using multi-criteria analysis methods in the DSS NooTron

Abstract

There were analyzed the possibilities of using multi-criteria analysis methods from the DSS NooTron (https://nootron.net.ua/) library considering system problems of a complex structure on a finite set of alternatives and criteria.The use of multi-criteria analysis methods is becoming an integral part of solving problems arising from the analysis, optimization, and evaluation of the effectiveness of complex weakly structured systems. These are such problems as comparative analysis and choice of the best alternative, making design decisions, vector optimization, resource allocation, diagnostics, rating compilation.Methods of multi-criteria decision analysis (MCDA) are used in many areas of science and practice. Quantitative methods of MCDA are of particular interest. These methods provide algorithms for dividing the system problem being studied into separate elements (decomposition), analyzing the selected blocks, determining the value of influence of each element on others (analysis), determining local (criterial) results and converting them into a global assessment (aggregation).The purpose of this work is to demonstrate the capabilities of multi-criteria analysis methods from the DSS NooTron library in system problems of a complex structure on a finite set of alternatives and criteria, including the tasks and methods implemented in the new version of NooTron DSS.A group of quantitative multi-criteria methods that are the most widely used and modified are chosen for the study, namely: the analytical hierarchy process, the analytical network process, the BOCR efficiency assessment methodology, the weighted sum method, the decision matrix method.The analysis of the structures of solved practical problems using the DSS NooTron led to the conclusion that this system provides a wide range of possibilities for system analytics, and also allows the analysis of weakly structured systems.The project “DSS NooTron” continues to evolve and improve. At the time of this writing, the following had been performed in the development of the system:1. The project architecture was improved on the basis of selected components of multi-criteria methods and React JavaScript framework for future support and scaling.2. Organized data exchange between components, their synchronization and processing of the application state.3. Improved interaction with the server to obtain intermediate results of solving the problem.4. A unified component flexible version of the analytical hierarchy process was developed using the React JavaScript framework.5. Implemented visualization of a dynamic hierarchical structure of a multi-criteria task in AHP.6. A simplified BOCR algorithm was developed to evaluate the effectiveness of IT projects.