Optimization of primal and dual network models of distribution

Abstract A direct `one pass' method of solution of the distribution problem is developed. Basis transformation is then applied to the original constraint equations of this to obtain a quadratic problem which can also be obtained by summation of finite element matrices in which the element constitutive parameter is respectively the route cost or inverse cost for the dual and primal problems. Using the method of steepest descent extrema for both these are obtained which are equal or close to the exact solutions.