Alternative crossover schemes

The results of the crossover experiments argue against a mechanism initiated by hy-drosilylation. An alternative mechanistic possibility, which is consistent with all of the data and which adequately accounts for the stereoselective formation of only diastere-omers 148 and 149, is given in Scheme 51. Complexation of the silane to an initially formed palladacycle such as 155 would afford a complex such as 156. Sigma bond metathesis to a mixture of 7r-allylpalladium complexes 157 (presumed major) and 158 (presumed minor) followed by reductive elimination from each of these would generate the observed products 148 (major) and 149 (minor). 

It is their wrter parallelism - i.e. the way in which crossover builds the solution - which makes genetic algorithms an attractive alternative to other optimization schemes such as simulating annealing. If trees of Os and Is are used instead of strings, it even becomes possible to breed computer programs by this method. 

---