Module cooperative 

Cooperative co-evolution — CCGA (Potter & De Jong 1994).

A high-dimensional problem is decomposed by assigning disjoint subsets of dimensions to two populations. Neither population holds a complete solution on its own: to score a member of population A, it is combined with a representative drawn from population B (and vice versa) to assemble a full-dimensional candidate, which the CoupledFitness then evaluates row-wise.

Where the coupling lives

Representative selection and full-genome assembly live in CooperativeCoEA::step — the algorithm owns the coupling, alongside RepresentativePolicy (in CooperativeCoEAParams). The CoupledFitness stays a pure, stateless row-wise objective (e.g. Rastrigin): it receives already-assembled full-dimensional populations and never performs selection or holds a lock.

Structs

CooperativeCoEA

Cooperative (CCGA-style) co-evolutionary algorithm.

CooperativeCoEAParams

Static parameters for CooperativeCoEA.

CooperativeState

Joint state for CooperativeCoEA: the shared CoEAState plus the per-population representative archives used by RepresentativePolicy::Archive.

Enums

RepresentativePolicy

Which member of the opposing population completes a partial candidate during fitness evaluation.