Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Issue N10 2017 year
In article the solution of a linear programming problem (LPP) in presence of uncertainty of parameters of target function and inequality restrictions by its reduction to the usual determined systems is offered. The algorithm of calculation of a modeled optimal problem is based on use of the simplex method which is built into the MATLAB system. The interval solution of a standard problem of optimization is represented as interval expansion of the optimal pointwise plan received on the basis of a nominal canonical set of equations. With use of interval representation of factors of the received nominal system the interval problem of an interior estimation of set of solutions of interval system of the linear algebraic equations (ISLAE) is solved and the interval vector (brus) corresponding to the optimal admissible plan is calculated. The admissibility of the discovered interval optimal plan is checked by its substitution in the pointwise sets of equations defining an exterior estimation of set of solutions ISLAE constructed on the basis of interval canonical system. For optimal angular matrices of the formalized pointwise systems their product on the discovered vector should not overstep the bounds of an interval vector of restrictions. The purpose of this research is an objective deriving of boundaries of the optimal target function, not exceeding the limit of possible deviations called by uncertainty of a problem data and conforming to inequality restrictions.