|List in two parts||6.3||127||2,580||538||406||1,050|
Table A.1 summarises the estimated effort, in terms of the number of trial solutions evaluated, required to solve (with at least 99% assurance) the problems presented in this book. Where problems were not solved a lower bound has been calculated based on assuming the very next run would have succeeded by generation 25.
The number of program executions required (for 99% probability of solving the problem) is estimated by multiplying the number of trial solutions by the mean number of times each was run during its fitness testing. Where the mean number of program executions per program tested is not available, the maximum is used to give an estimated upper bound. Run time reductions via: ancestor fitness re-use (cf. Section D.5), ADF caching (cf. Section D.6) and avoiding fitness evaluation of individuals produced by reproduction, are excluded.
Not all of Genetic Programming and Data Structures is online.