Algorithm of solving corroding construction optimization problems based on flexible tolerance method

Abstract

In current paper authors formulated a new problem of hinged-rod constructions optimal design, which considers physicochemical processes in construction elements that cause reduction of their bearing capacity and assumes that solution is obtained with a given accuracy. The search for an optimal solution is made on discrete non-metrical space of varied parameters.

Actuality of the problem of corroding constructions optimal design is determined by the requirements for their high reliability and minimal consumption of materials. Utilization of existing algorithms for optimal design of such constructions allows to achieve required solution accuracy only with high computational cost. The paper describes creation of effective algorithm based on flexible tolerance method which allows to obtain solution with given accuracy and, therefore, to ensure required level of reliability of designed construction. Optimization algorithm uses neural network module of computational error control, which allows to change parameters of numerical solution of differential equation system modeling the influence of aggressive environment while solving the optimization problem. It allows to reduce computational cost at initial stages of search for the solution and to ensure required solution accuracy in the vicinity of the extremum.

Analysis of results of numerical experiments allows to make a conclusion about high performance of optimization algorithm while ensuring given accuracy of problem solution. Utilization of developed algorithm will allow to solve the problems of  corroding hinged-rod constructions optimal design.