Reducing the number of unknowns in a constrained minimisation problem-an appliation to material balances

The mineral processing material balance problem and the available mainframe packages for automated solution are outlined. In mathematical terms the material balance problem is a constrained minimisation problem. An efficient solution approach is developed which reduces this to an unconstrained minimisation problem with fewer variables. The computer storage space of this apporach is much smaller than previous approaches to automated material balances. It is thus able to run on a personal computer and compete against mainframe packages. A prototype of this approach is developed, and test results on a large problem are included. By developing a tailor-made package for personal computers, it is hoped that a more widespread use of material balance packages will be made.