1 1.0
.5
0.8
.02
P3(i n) 0.6
0.4
0.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
i
P3(i I n) at n = (2000010001002000)
Figure 5-6
5.5 Converged Limiting Stationary Distributions
The following data represent converged three-operator stationary distribution
results for both four and five-bit problems at a variety of population sizes. The results
recorded in Figures 5-7 through 5-16 are products of the computer program
GET_3STAT.F included in Appendix D. They are obtained by repeatedly multiplying a
current state probability vector by the three-operator state transition matrix until a termi-
nation criterion representing approximate convergence is attained. The starting probabil-
ity vector is the multinomial distribution corresponding to a uniformly distributed P3(i I n)
array, and the termination criterion is that the sum of the probabilities for all nonuniform
population states is less than 0.004.
All of the results reported here are for extremely small a (approaching zero) and
thus, as predicted by the model, only the states corresponding to uniform populations