mixsqp release published on CRAN (2020-01-29)

This version of the package contains various improvements and bug fixes that should result in more reliable and faster convergence. Some of the new key features include automatic normalization of the data rows for more numerically stable behaviour, and an option to compute a low-rank approximation of the data matrix, which provides more efficient gradient and Hessian computations when the number of columns is large.