Optimization method with parallel computations

An optimization method with parallel computations, the method comprising:

• performing multiple stages of parallel calculation of values of a target function defined by P independent parameters using computing resources of one or more graphics processors (GPU),
• wherein an entire one-dimensional array of the calculated values of the target function of length
• ∏
• j
• =
• 1
• P
• ⁢
• W
• j
• that needs to be computed is divided into groups of size η that are calculated in parallel at the
• L
• =
• ∏
• j
• =
• 1
• η
• ⁢
• W
• j
• η parameter points;
• wherein a number of simultaneously calculated parameters η in each group and a number of calculation points Wj of the target function at a given interval Dj for each j-th desired parameter in the group is selected based on a possible number of parallel calculations R=G·M·T,
• where G is a number of GPUs, M is a number of cores in each GPU, and T is a number of threads in each core;
• and
• outputting the calculated P parameters that correspond to the global extremum of the target function, and
• wherein a full cycle of calculating points is carried out for the number of consecutive iterations, defined as integer division
• ⌈
• L
• R
• ⌉
• with rounding up.