What is New in What'sBest! 13.0

Release 13 of What'sBest! includes a wide range of performance enhancements and new features:

Faster Solutions on Linear Models with Improved Simplex Solver

• Simplex LP algorithm improvements boost speed and robustness. Primal Simplex Solver averages 90% faster and Dual Simplex Solver averages 45% faster.

Better Performance on Integer Models

• Improved solution times across broad classes of IP models from enhancements due to improved knapsack related cuts, superior default node selection rules, and branching variable selection rule options.

Expanded Capabilities of Efficient Solution of Quadratic Problems

• If you work with covariance matrices, you'll appreciate the new Positive Definite (POSD) feature. If estimating a covariance matrix for a portfolio, a constraint may be added to force the matrix to be positive semi-definite (a property required of any covariance matrix).  New reformulation capabilities give improved performance on quadratic portfolio models with semi-continuous variables (e.g. min-buy quantities), and/or cardinality constraints limiting the number of assets in a portfolio.

Smarter Presolver

• New preprocessing for linear programs significantly reduces coefficient density of certain dense matrices. This can significantly improve performance on models such as production planning with many periods.

Improved Nonlinear Solver

• Upgraded default settings provide performance improvements overall on nonlinear models. Faster processing of long (1000’s of terms) nonlinear expressions in nonlinear models.

New Functionalities

• Support for additional math and probability functions, e.g., Erlang B and C formulae. Add a feature to generate a K-Best report when generating K best solutions.
  

© Copyright 2015 Lindo Systems Inc.