Optimization procedure

For our optimizations, we use home-made software routines calculating the selection gain and its standard deviation for the considered scenarios.We developed a package called "selectiongain", which runs under the open source statistic software R. For more information, you can also look at the publication of Xuefei Mi. Additionally, we offer lectures and training courses on this software tool.

The optimization is roughly done like that. We firstly define the breeding scheme, the budget, and the species via its specifique variance components. For the cost functions we need rough estimates for all steps in a breeding program, for instance for crossing, selfing, DH production, nursery rows, seed multiplication and field plots. Finally the grid of possible allocations must be defined, that is the minimum and maximum number of lines per cross, test locations, and testers, the kind of testers, and so on. We then optimize for each breeding strategy the allocation of resources and finally compare alternative optimized breeding strategies. For more details, please visit my YouTube channel.