Author:
,Polishchuk O.D.,Yadzhak M.S.,
Abstract
The methodology for interactive evaluation of the functioning efficiency of the motor transport system for a large city based on the use of U -statistics methods is formalized. To optimize this technique, we propose efficient algorithmic constructions for parallel execution of local, aggregated, and forecasting evaluation of system components on modern computers, namely, multi-core computers, clusters, hybrid architectures, and high-performance computing environments. The results obtained allow for real-time evaluation of the efficiency for the city’s motor transport system. Keywords: motor transport system, U -statistics, evaluation, aggregation, forecasting, parallelisation of computations, speed up, avtonomous branches.
Publisher
V.M. Glushkov Institute of Cybernetics