The use of genetic algorithm to evaluate objects based on partially contradictory data

Robert Mayer

Abstract


The problem of obtaining optimal estimates based on redundant and contradictory data provided by one or more experts is considered. As an example, the evaluation of the didactic complexity of physical concepts is discussed. The data contains: 1) value judgments such as: "the complexity of the concept C is approximately equal to the sum of the complexities of the concepts A, B multiplied by k, and the number b" (b is an integer, k=1; 1,1; 1,3; 1,5); 2) confidence coefficients in the truth of judgments from the interval [0; 1]. There are cases when the complexities of concepts are almost the same (k = 1), differ slightly (k = 1.1), noticeably (k = 1.3), significantly (k = 1.5) and very strongly (k > 2). A computer program based on ABCPascal is proposed that implements a genetic algorithm that simulates the evolutionary process using mixing, mutation, and natural selection operators. They are applied to chromosomes consisting of genes, which are estimates of objects. To solve the problem, we need to construct an objective function that calculates the sum of the modules or squares of all discrepancies. The program finds the values of the estimates x[j] (that is, the genes that make up the chromosomes) for which the target function is maximal. The optimization results of expert assessments of the didactic complexity of educational concepts, which contain contradictory data, are analyzed

Full Text:

PDF (Russian)

References


Burakov M. V. Geneticheskij algoritm: teorija i praktika. – Sankt-Peterburg: Sankt-Peterburgskij gosudarstvennyj universitet ajerokosmicheskogo priborostroenija, 2008. – 164 s.

Gladkov L. A. Geneticheskie algoritmy: uchebnik / L. A. Gladkov, V. V. Kurejchik, V. M. Kurejchik. – Moskva: OOO Izdatel'skaja firma "Fiziko-matematicheskaja literatura", 2010. – 366 s.

Velichkovskij B.M. Kognitivnaja nauka: Osnovy psihologii poznanija: v 2 t. T. 1. M.: Smysl: Akademija, 2006. – 448 s.

Egorov S. A. Metod primenenija teorii nechetkih mnozhestv dlja ocenki processa obuchenija // Vestnik nauki i obrazovanija. 2020. 11–2 (89). – S. 27–33.

Zaginajlo M. V. Geneticheskij algoritm kak jeffektivnyj instrument jevoljucionnyh algoritmov / M. V. Zaginajlo, V. A. Fathi // Innovacii. Nauka. Obrazovanie. – 2020. # 22. – S. 513-518.

Zade L.A. Razmytye mnozhestva i ih primenenija v raspoznavanii obrazov i klaster-analize. – M.: Mir, 1980. – 390 s.

Ivshina G. V., Mardanov M. V. Nechetkoe modelirovanie rezul'tatov metoda jekspertnyh ocenok v pedagogicheskom jeksperimente // Obrazovanie i samorazvitie. – 2007. # 2 (4). – S. 13 – 19.

Majer R. V. Slozhnost' uchebnyh ponjatij i tekstov: monografija. – Glazov: GIPU, 2024. – 132 s.

Mitin A. I., Filicheva T. A. Ocenka kachestva obrazovatel'nyh uslug: modelirovanie na baze teorii nechetkih mnozhestv i nechetkoj logiki // Modelirovanie i analiz dannyh. – 2016. # 1. – S. 3–20.

Poleshhuk O. M. O primenenii nechetkih mnozhestv v zadachah postroenija urovnevyh gradacij. Lesnoj vestnik. – 4(13). 2000. – S. 143-146.

Pospelov D. A. Situacionnoe upravlenie: teorija i praktika. – M.: Nauka. – Gl. red. fiz.- mat. lit., 1986. – 288 s.

Subetto A. I. Komp'juternaja kvalimetrija v obrazovanii. Perspektivy razvitija (nauchnyj obzor) // Kachestvo v obrabotke materialov. #1 (5). 2016.

Subetto A. I. Vvedenie v kvalimetriju vysshej shkoly (knigi I - IV). M.: Issledovatel'skij centr Gosobrazovanija, 1991.

Sutuzhko V. V. Obshhenauchnye aspekty teorii ocenki // Vestn. Volgogr. gos. un-ta. Ser. 7, Filosofija. 2009. # 1 (9). S. 42 – 46.

Flegontov A. V. Mjagkie znanija i nechetkaja sistemologija gumanitarnyh oblastej / A.V. Flegontov, V.A. Djuk, I.K. Fomina // Programmnye produkty i sistemy. 2008. # 3. – S. 97–102.


Refbacks

  • There are currently no refbacks.


Abava  Кибербезопасность ИБП для ЦОД СНЭ

ISSN: 2307-8162