Integrated into the Wolfram Language is a full range of state-of-the-art local and global optimization techniques, both numeric and symbolic, including constrained nonlinear optimization, interior point methods and integer programming—as well as original symbolic methods. The Wolfram Language's symbolic architecture provides seamless access to industrial-strength system and model optimization, efficiently handling million-variable linear programming, and multithousand-variable nonlinear problems.
FindFit — optimal nonlinear unconstrained or constrained fit to data
LinearProgramming — real and integer linear programming in matrix form
LeastSquares — least-squares problem in matrix form
FindShortestTour — solve a traveling salesman problem