An evolutionary search strategy utilizing two normal distributions to gener
ate children is presented. This Bell-Curve Based (BCB) evolutionary algorit
hm is similar in spirit to (mu+mu) evolutionary strategies but with fewer p
arameters to adjust. Extensive tests regarding the sensitivity of BCB param
eters to performance are provided. The test suite includes continuous varia
ble constrained hub design problems, mixed discrete and continuous variable
constrained hub design problems, and an unconstrained highly multimodal di
screte optimization problem.