A comparative study of linear complementarity programming models and linear programming models in multi-regional investment analysis
T. Takayama · 1984
This paper compares the linear complementarity programming (LCP) investment model with the linear programming (LP) investment model, using the bauxite-alumina-aluminum industry as an example. The differences in results obtained from the models are examined with respect to two criteria: (i) the stability of solutions, given perturbations in parameters; and (ii) the ability to handle constraints on the supply of inputs. Under the first, three parameter changes are tested - changes in electricity costs, changes in transport costs, and changes in the tax on bauxite supply; under the second, two examples are studied - where policy decisions constrain the quantity of bauxite production and where the supply of electricity for smelting within a region is limited by existing generation capacity. The major conclusion is that, although the basic structure of the LCP and LP models is very similar, the LCP model is capable of dealing with a wider range of policy questions. Development of LCP software is strongly urged in order to take full advantage of this technique.