Parallel Solver for Semidefinite Programming

A location problem to find the minimum-cost assignment of given plants to candidate locations is a typical combinatorial optimization problem in logistics. When the number of plants increases, however, the computational cost of the location problem grows up extremely. If we enumerated all possibilities, the cost would be proportional to the factorial of the number of plants. In general, meta-heuristic techniques, for example, Genetic Algorithm, are employed to find an approximate solution. Although meta-heuristic techniques are very useful, they cannot assure the accuracy of the solution, because they search only feasible solutions, hence their objective values are always only upper bound