Kajsa Ljungberg's home page.
## QTL-mapping softwareSoftware developed as part of thesis work. |

## Evaluating the LS objective functionThe objective function is the residual norm of a LS problem min ||Xb-y||, where X depends on the position in the search space. However some columns of X are constant. For efficiency we use updated QR factorizations and do not solve the complete problem since only the residual is needed, not the regression parameters b.The updating is efficient only if the number of constant columns in X is large enough, otherwise a library QR factorization should be used. The complete LS problem should still not be solved.
## Global optimizationWe have used the DIRECT global optimization algorithm to find point in search space with minimal value of objective function. The algorithm can be coupled with any objective function routine fulfilling given specifications.DIRECT original reference: D. Jones, C.D. Perttunen and B.E. Stuckman. Lipschitzian optimization without the Lipschitz constant. Journal of Optimization Theory and Application, Vol 79, pp. 157-181, 1993. Retrieving the source code below implies that you have agreed to cite in any publication or presentation where the algorithm has been used for the work. Source Readme The project is a collaboration between Kajsa Ljungberg and Sverker Holmgren at the Department of Scientific Computing, Örjan Carlborg, Kateryna Mishchenko at Mälardalen university college , Leif Andersson at the Department of animal breeding and genetics at SLU, Martina Persson from the Department of Mathematics, Mathematical Statistics Group at Uppsala University, and Razaw AL-Sarraj and Dietrich von Rosen from the Department of biometrics and informatics at SLU. |