banner

Simplex Method

  • Used for linear optimization problems
  • Fastest and most reliable for linear models
  • Guaranteed to find the global optimum for linear problems
  • Only works when the Assume Linear Model box is checked1

GRG (Generalized Reduced Gradient) Nonlinear Method

  • Used for nonlinear optimization problems
  • Can handle smooth, continuous nonlinear functions
  • May find local optima rather than global optima
  • Slower than Simplex but faster than Evolutionary
  • Default method when Assume Linear Model is unchecked1
  • Works well for problems with continuous derivatives

Evolutionary Method

  • Used for non-smooth or discontinuous problems
  • Can handle discrete variables and non-smooth functions
  • Slowest of the three methods
  • May not find the global optimum, but can often find good solutions for complex problems
  • Useful when other methods fail due to discontinuities or discrete variables

Key Considerations

  • Problem Type: Linear problems should use Simplex. Nonlinear but smooth problems can use GRG Nonlinear. Complex or discontinuous problems may require Evolutionary.
  • Solution Quality: Simplex guarantees global optima for linear problems. GRG may find local optima. Evolutionary is less precise but can handle more complex problems.
  • Computation Time: Simplex is fastest, followed by GRG, with Evolutionary typically being the slowest.
  • Starting Points: GRG and Evolutionary methods can benefit from multiple starting points to increase the chance of finding the global optimum1.
  • Constraints: All methods can handle constraints, but their effectiveness may vary depending on the problem structure.

To get the best results, it’s often recommended to:

  1. Use Simplex for linear problems
  2. Try GRG Nonlinear first for smooth nonlinear problems
  3. Use Evolutionary as a last resort for complex, discontinuous problems
  4. Experiment with different starting points, especially for GRG Nonlinear1
  5. Consider using the MultiStart option with GRG Nonlinear for global optimization2

By understanding the strengths and limitations of each method, you can choose the most appropriate solver for your specific optimization problem in Excel.

Additional Learning Resource to check out

Some Solver resources Frontier Solver They have a lot of template files with case studies to help learning

Microsoft Offical Documentation

Microsoft Offical Tutorial

Open Solver Alternative solver solution with advanced models.