where o is a constant, the main function of o is to ensure
the positive result to fit growth model of TA, O can be obtained
by fuzzifying the crisp value of demand in each period
according to expression (1), Wi is the weight of the
evaluation factor i and di is the value of fuzzy variable
belonging to evaluation factor i.
Layer 3 is the transfer layer. The connection weights in
this layer are obtained by the fuzzy AHP method mentioned
earlier in Section 3. There is only one neuron in this
layer, whose value is defined in expression (3).