Genetic Algorithm and Direct Search Toolbox™ 2.3

Product Description

• Introduction and Key Features
• Graphical User Interface and Command-Line Functions
• Genetic Algorithm Tools
• Direct Search and Simulated Annealing Tools
• Solving Constrained Optimization Problems and Using Other Solvers
• Displaying, Monitoring, and Outputting Results

Solving Constrained Optimization Problems and Using Other Solvers

The genetic and direct search algorithms implemented in the toolbox let you solve optimization problems with nonlinear, linear, and bound constraints. For linearly constrained optimization problems, the algorithms identify active linear constraints and bounds to generate search directions, or mutants for the genetic algorithm. For nonlinearly constrained problems, the algorithms formulate a subproblem subject to linear constraints and bounds using penalty and Lagrange parameters. They find an approximate solution to the subproblem, update the penalty and Lagrange parameters to formulate a new subproblem, and iterate until it is solved within a specified accuracy.

Output showing solutions reached when using only the genetic algorithm and when using the genetic algorithm with a function from the Optimization Toolbox. Combining algorithms can produce more accurate results. Click on image to see enlarged view.

Using Other Functions and Solvers

The Genetic Algorithm and Direct Search Toolbox is closely integrated with MATLAB and the Optimization Toolbox. You can use the genetic algorithm and pattern search to find good starting points and then use the Optimization Toolbox solvers or MATLAB routines to further refine your optimization. Solvers are available for both constrained and unconstrained optimization problems.

By combining algorithms, you can leverage the strengths of the toolboxes and MATLAB to improve the quality of your solutions.