A fisheries model with random effects 


Model descriptionThe catchage model from Schnute and Richards (1995) has been used compare the computational speed of AD Model Builder with other statistical software packages (see here). It is conceptually appealing to assume that the annual recruitements are random effects, i.e. they are drawn from a common normal distribution. The random effects module then lets you estimate the mean and the variance of the normal distribution.The model is memory consuming and requires the following command line options in order to run quickly: caest est l1 30000000 l2 100000000 l3 100000000 nl1 10000000 nl2 10000000 nl3 10000000 or it can be run a bit slower using less memory using the ndb optiion to calculate the derivatives in three blocks caest est ndb 3 l1 10000000 l2 70000000 l3 10000000 nl1 20000 00 nl3 2000000 or if you really don't have much memory available try ndb 10 option caest est ndb 10 l1 10000000 l2 10000000 l3 3000000 nl1 1000000 nl3 1000000 The computational time is 31s on a 2006pc running linux. 