CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
GENETIC ALGORITHM OF SELECTING INFORMATION SYSTEMS COMPONENTS BASED ON EXPERIMENTAL EVALUATIONS OF QUALITY CRITERIA
 |
Read | Gusev Aleksandr A., Ilin Dmitry Yu., Nikulchev Evgeny V. GENETIC ALGORITHM OF SELECTING INFORMATION SYSTEMS COMPONENTS BASED ON EXPERIMENTAL EVALUATIONS OF QUALITY CRITERIA // Caspian journal : management and high technologies. — 2019. — №2. — pp. 113-125. |
Gusev Aleksandr A. - post-graduate student, https://orcid.org/0000-0003-2437-8537, https://elibrary.ru/author_items. aspauthorid=835966, Kuban State University, alexandrgsv@gmail.com
Ilin Dmitry Yu. - Post-graduate student, http://orcid.org/0000-0002-0241-2733, https://elibrary.ru/author_items.asp? authorid=892115, MIREA - Russian Technological University, i@dmitryilin.com
Nikulchev Evgeny V. - Doct. Sci. (Engineering), Professor, Professor of the Chair of Systems Control and Modelling, http://orcid.org/0000-0003-1254-9132, https://elibrary.ru/author_items.asp?authorid=396636, MIREA - Russian Technological University, nikulchev@mail.ru
The article presents a methodology for assessing the effectiveness of a set of software components based on repro-ducible experiments under the control of a genetic algorithm. To represent the sets of components in the form of natural gen-otypes, an encoding mapping and reverse mapping is introduced to decipher the genotype. In the first step of the technique, the genetic algorithm creates an initial population of random genotypes, which are then converted into estimated sets of software components. Then, each set is initialized and a given list of operations is executed with experimental measurements taken according to 14 specified individual performance criteria. Based on experimental measurements, taking into account weighting factors that set goals in the field of Quality of Service, the genetic algorithm calculates the integral quality functional for each set of components of the information system under investigation, after which genetic operators are performed and the generation of improved genotypes is generated and experimental measurements are done for the corresponding sets, the procedure is repeated until the stop conditions are met. The article shows the application of the proposed methodology to evaluating the effectiveness of selecting Node.js components, for which a MATLAB genetic search program was developed as well as the experiment scenario for a virtual machine running the Ubuntu 16.04 LTS operating system. The virtual machine was deployed using the Vagrant virtual environment configuration tool.
Key words: качество систем и программ, эффективность взаимодействия программ, генетический алгоритм, эволюционные вычисления, вычислительные эксперименты, информационная система, quality of systems and programs, the effectiveness of program interaction, genetic algo