![Scientific and technical journal «Priborostroenie»](/images/mag-pr.png)
Journal
Scientific and technical journal «Priborostroenie»
UDK519.687
Issue:2 (63)
Download PDF350 Kbyte
The concept of heterogeneous workflows is formulated and following computational modes are identified: batch, streaming, and iterative. The necessity of developing of scheduling algorithms for each of modes is determined. Developed scheduling algorithms are based on machine learning methods, evolu-tionary approaches, and artificial intelligence methods. A technology has been developed that ensures the collaborative execution of heterogeneous workflows based on a cascade of developed optimization meth-ods.