Maximum likelihood estimates for multinomial probabilities via geometric programming

SUMMARY This paper illustrates how the duality theory of geometric programming and the, special methods available for solution can be used to obtain maximum likelihood estimates for multinomial probabilities. Three examples are given.