Trees in available wild forest10 stands j =

1,

c, 300 have been measured in terms of desirable traits i = 1,

c, 12 such as rate of growth, disease resistance, and wood density. Computations of averages and standard deviations in traits over the stands then produced numbers zi, j of standard deviations that stand j falls above or below the average for trait i. Corresponding scores for the offspring of a planned breeding program may be assumed to combine linearly (i.e., the result will be a sum of scores for the wild stands used, weighted by the proportion each stand makes up of the breeding population).

(a) Formulate a multiobjective LP to choose a breeding population that maximizes the offspring standard deviation score of each trait.

(b) Convert your model to a goal program with a target of breeding offspring 2 standard deviations above the mean on all traits. Weight the undersatisfaction of each goal equally.