Meta-Genetic Programming: Co-evolving the Operators of Variation - Bruce Edmonds

8. Further Research

• thorough trials of MGP over a variety of problem domains;

• the optimal tuning of MGP algorithms;

• the biases introduced by the language the operators are expressed in;

• a comparison of MGP and ADF/ADM techniques;

• the merits of combining MGP and ADF/ADM techniques;

• the effectiveness of recursive genetic programming (as illustrated in figure 16) where an evolving population of operators act upon itself.

Generated with CERN WebMaker