MathWorks today announced the addition of mixed-integer linear programming (MILP) to MATLAB. Now available as part of Optimization Toolbox with Release 2014a, this new solver gives users the ability to solve optimization problems that require integer solutions, such as decisions on the number of stocks to buy or sell.
For problems requiring integer solutions, algorithms that employ integer programming techniques enable companies to make optimal decisions. Tools based on MILP can result in significant financial gains and savings in applications such as portfolio optimization and resource allocation. The new solver can be used with MATLAB deployment products to create standalone applications based on MILP and to integrate algorithms that use MILP with other languages such as Java and .NET.
Many business problems require MILP algorithms to find answers that are integer numbers. For example, variables that represent stock shares to purchase must be integer values to execute a trade. Similarly, variables that represent the on/off state of power generators require binary values (0 or 1). Instead of rounding the solution from a traditional continuous solver, which often violates problem constraints, a MILP solver finds the optimal integer solution.
"Analysts and engineers use MILP to find the optimal solutions to common business problems such as portfolio optimization, resource allocation, and scheduling," said Seth DeLand, technical marketing manager, MathWorks. "By including the mixed-integer linear programming solver in Optimization Toolbox, MathWorks is enabling users to build and deploy decision support systems based on MILP that can be used throughout the enterprise."
Optimization Toolbox with MILP solver is available immediately in Release 2014a (R2014a). For more information, see R2014a Release Highlights.