Goal programming besides, being a broadly used decision tool, is shown in this paper to be a powerful forecasting device too. It is used as the main component of an energy demand forecasting system, performing a function of an integrator of information on the future growth of energy demand. This information comes from a variety of sources. The most important is a set of econometric models, based on different approaches, which explain past behaviour of the energy demand system. Other sources of information used are statements on government energy policy, feasibility or technical reports and, finally, the forecaster's guess. Thus GP is intended to relax many of the assumptions of conventional econometric forecasting, which render it inadequate under the prevailing conditions of quickly changing energy structures. An implementation of the system on the Greek economy is also described in this paper. Four scenarios of energy demand growth up to 1990 are studied. The levels of energy requirements in these scenarios vary as a consequence of the patterns of economic development and rates of increase in the price of oil.
Read full abstract