DEPENDENT-CHANCE GOAL PROGRAMMING AND ITS GENETIC ALGORITHM-BASED APPROACH

This paper develops a general formulation of dependent-chance goal pro gramming (DCGP) which is an extension of stochastic goal programming i n a complex stochastic system, and gives an example of water allocatio n and supply to show the application of DCGP. A genetic algorithm base d approach is also presented to solve such a model. DCGP is available to the systems in which there are multiple stochastic inputs and multi ple outputs with their own reliability levels. The characteristic of D CGP is that the chances of some probabilistic goals are dependent, i.e ., the goals cannot be considered in isolation or converted to their d eterministic equivalents. Finally, Monte Carlo simulation is also disc ussed for calculating the chance functions in complex stochastic const raints.