Capacity expansion problems based on fuzzy dynamic programming

This paper discusses capacity expansion problems for telecommunication network facilities, based on fuzzy dynamic programming. Although cost functions, discount rates, demand functions, and so on should be given, they usually cannot be defined clearly because of technical developments or business fluctuations. This paper represents undefined factors by fuzzy numbers with triangular membership functions (TFN). Multiplication or division of TFN does not give rigid TFN, but we approximate them to TFN for ease of calculation. Three methods based on this approximation are compared, using numerical examples. Approximation accuracies are confirmed by strict calculation using removal of defining orders of fuzzy numbers.