No abstract

PArAllel cOmPuTATIOnAl AlGOrIThmS In ThermAl PrOceSSeS In meTAllurGY AnD mInInGPurpose. Formation of parallel algorithms in the thermal process simulation in metallurgy and mining. The proposed parallel form of the algorithms must be maximal, and, therefore, have the minimum possible implementation time in parallel computing systems. Elimination of recurrent computation structure of the desired decision vectors, which, as a rule, leads to the rounding errors accumulation. Such class simulation problems are realized by multiprocessor computing systems.methodology. Implementation of parallelizing process of mathematical problem definition is realized by an approach based on the "odd-even" reduction algorithm. The essence of this approach lies in exclusion of simulation coefficients of the process under research by realizing elementary rows transformations of the constructed equations system. A directly parallel form of the algorithm for solving problems is realized by a numerical-analytical approach. It is shown that the compiled parallel form is the maximum, which, in turn, provides minimum time for solving the set problems by multiprocessor computing systems.Findings. The presented research studies in this paper showed a high efficiency of parallelization of systems of tridiagonal structure linear algebraic equations by example of solving thermal problems. The proposed numericalanalytical method for parallelizing tridiagonal systems does not impose any restrictions on the grid nodes topology of the computational domain. With respect to parallel computations of arithmetic expressions, the original data error is separated from the rounding operations by the proposed method. This approach excludes the recurrent structure of the desired decision vectors computation, which, as a rule, leads to accumulation of rounding errors. The proposed parallel form of the algorithms must be maximal, and, therefore, have the minimum possible implementation time in parallel computing systems. Computational experiments conducted by a multiprocessor computer system showed high efficiency of the developed parallel algorithms.Originality. Within decomposition algorithms, based on the "odd-even" reduction method, a new approach to the distributed solution of linear algebraic equation systems is proposed for the first time, which differs from the known methods in the closed parallel form with respect to the central grid node and with a high degree of vectorization. There was proposed, analyzed and implemented a new approach to the solution of metallurgical production problems, which allows increasing economy, productivity and speed of computations. It is proved that this approach provides the highest computation vectorization degree, predetermines the maximum parallel algorithmic form and, as a consequence, the minimum possible time for implementing algorithms on parallel computing systems.Practical value. By using a high-performance multiprocessor system, the developed approach allows processing and interpreting the thermal experiments r...

Context. The problem of increasing the efficiency of optimization methods by synthesizing metaheuristics is considered. The object of the research is the process of finding a solution to optimization problems. Objective. The goal of the work is to increase the efficiency of searching for a quasi-optimal solution at the expense of a metaheuristic method based on the synthesis of clonal selection and annealing simulation algorithms. Method. The proposed optimization method improves the clonal selection algorithm by dynamically changing based on the annealing simulation algorithm of the mutation step, the mutation probability, the number of potential solutions to be replaced. This reduces the risk of hitting the local optimum through extensive exploration of the search space at the initial iterations and guarantees convergence due to the focus of the search at the final iterations. The proposed optimization method makes it possible to find a conditional minimum through a dynamic penalty function, the value of which increases with increasing iteration number. The proposed optimization method admits non-binary potential solutions in the mutation operator by using the standard normal distribution instead of the uniform distribution. Results. The proposed optimization method was programmatically implemented using the CUDA parallel processing technology and studied for the problem of finding the conditional minimum of a function, the optimal separation problem of a discrete set, the traveling salesman problem, the backpack problem on their corresponding problem-oriented databases. The results obtained allowed to investigate the dependence of the parameter values on the probability of mutation. Conclusions. The conducted experiments have confirmed the performance of the proposed method and allow us to recommend it for use in practice in solving optimization problems. Prospects for further research are to create intelligent parallel and distributed computer systems for general and special purposes, which use the proposed method for problems of numerical and combinatorial optimization, machine learning and pattern recognition, forecast.

No abstract

No abstract