Neural Network Modelling of Constrained Spatial Interaction Flows



where δ is the step size that has to be chosen a priori and ε an uniformly distributed
random value with
ε [0,1]. The probability of change of the parameter is calculated
as

pk ( n ) = (1+exp ( Ck (n ) / T (n ) ) )1                                               (29)

with Ck (n) given by the correlation

Ck (n) = [wk (n -1) - wk (n - 2)] [Q (x, У, wk (n -1)) - Q (x, У, wk (n - 2))]     (30)

= [wk (n)] [∆Q (χ, y, wk (n))]

The weight will be incremented in a given fixed magnitude δ, when Δwk > 0, and the
opposite when it is less than zero. The sign of
Ck indicates whether Q varies in the
same way as
wk. If Ck0, both Q and wk will be raised or lowered. If Ck < 0, one
will be lowered and the other one raised.

If T is too small, the algorithm gets trapped into local minima of Q. Thus, the value of T
for each iteration, T(n), is chosen using the following heuristic annealing schedule:
where 3
H denotes the number of weights. The annealing schedule controls the
randomness of the algorithm. When
T is small, the probability of changing the
parameters is around zero if
Ck is negative and around one if Ck is positive. If T is
large, then
pk0.5. This means that there is the same probability to increment or
decrement the weights and that the direction of the steps is now random. In other
words, high values of
T imply a random walk, while low values cause a better
correlation guidance (see Bia 2000). The effectiveness of Alopex in locating global
minima and its speed of convergence critically depends on the balance of the size of the
feedback term
Δwk ΔQ and the temperature T. If T is very large compared to Δwk ΔQ

δ

3 HN


___ n1

Σ Σ ICk (n')
k n’=n N


T ( n -1)


if nis a multiple of N
otherwise


(31)


18



More intriguing information

1. The name is absent
2. Regional dynamics in mountain areas and the need for integrated policies
3. The name is absent
4. The name is absent
5. Are Public Investment Efficient in Creating Capital Stocks in Developing Countries?
6. Financial Market Volatility and Primary Placements
7. Can we design a market for competitive health insurance? CHERE Discussion Paper No 53
8. LOCAL PROGRAMS AND ACTIVITIES TO HELP FARM PEOPLE ADJUST
9. Improving Business Cycle Forecasts’ Accuracy - What Can We Learn from Past Errors?
10. The name is absent