A computational procedure based on the results of Barankin and Dorfman [1], for minimising a convex quadratic function subject to linear constraints is developed. The applicability of the proposed procedure for linear programming problems is indicated.