BIGDOT - Large Scale Optimization Software Library

BIGDOT is a gradient-based optimization library, similar to DOT. Users link the BIGDOT library with their program to perform optimization. BIGDOT is intended to solve very large scale optimization problems and can handle linear, non linear, constrained and unconstrained optimization problems. BIGDOT is capable of solving optimization problems with hundreds of thousands of design variables and can handle problems with large numbers of active constraints, up to the number of design variables. In addition discrete/integer and mixed variable problems can also be solved. For discrete variable problems, there is no guarantee of obtaining the true minimum. The techniques in BIGDOT are intended to efficiently obtain a near optimum discrete solution for large scale optimization problems.