Розробка методу класифікації бойових колісних машин із застосуванням нечіткого кластерного аналізу

The object of this study is combat wheeled vehicles (CWVs). The task addressed is the categorization of CWVs. The devised CWV categorization method is based on the use of the fuzzy c-means (FCM) algorithm, which determines the centers of fuzzy clusters and their corresponding functions and memberships, which can take values in the interval from 0 to 1. Therefore, the degrees of membership of CWV samples to fuzzy clusters have been determined, which together define the fuzzy division of the initial set of CWV samples. The minimum number of samples required to solve the fuzzy clustering problem by determining the values of the objective function and the magnitude of its increment per sample with a sequential increase in the number of samples is 55 pieces. It has been confirmed that the maximum number of clusters at level 6 satisfies the needs of categorization and does not require their increase due to the presence of individual CWV samples with large degrees of membership in 2-clusters. It has been proven that with a weight parameter value of 1.68, the fuzziness of the membership matrix ensures an average level of membership of samples to 6 clusters at a level of not less than 99 %. The proposed CWV categorization method establishes a correspondence between the technical characteristics of the samples and their functional purpose. This makes it possible to take into account the uncertainties caused by the assignment of samples with intermediate characteristics between groups to one group. The resulting categorization results establish benchmarks to which CWV samples should approach when designing CWV types, which are constructed on the basis of unified units and assemblies. The results of this study could be used to determine CWV samples of the same type under conditions of a significant variety of options for providing units