![[Home]](images/bhome.gif){kind=small}
![[What's New]](images/bnews.gif){kind=small}
![[Products & Services]](images/bprdsrv.gif){kind=small}
![[Contents]](images/btoc.gif){kind=small}
![[Feedback]](images/bfeed.gif){kind=small}
![[Search]](images/bsrch.gif){kind=small}
To download a free Evaluation Version of the Large-Scale GRG Solver Engine Version 3.5, click here to sign up (free of charge) for Frontline's Private Web for Solver Users. Then click on Download, scroll to the bottom and follow the link to Download Evaluation Version of Large-Scale GRG Solver Engine.
![[Dividing Line Image]](div.gif){kind=small}
Frontline Systems is excited to offer the Large-Scale GRG Solver Engine Version 3.5 for Microsoft Excel 2000 and Excel 97. The Large-Scale GRG Solver is a new field-installable Solver Engine that "plugs into" the Premium Solver Platform. It extends the power of the Premium Solver Platform to solve smooth nonlinear optimization problems of up to 4,000 variables and 4,000 constraints. It also supports integer and binary integer variables, and the new "alldifferent" constraint introduced in the Premium Solver Platform V3.5.
The Premium Solver Platform is 100% upwardly compatible with the standard Excel Solver. It replaces the standard Solver and appears when you select the usual Tools Solver... menu choice. You'll see a familiar Solver Parameters and Solver Options dialog with many new features. Your existing models can be solved with no changes, faster and more easily than before -- with any appropriate Solver engine for the Premium Solver Platform, including the Large-Scale GRG Solver Engine! And your VBA code designed to control the standard Solver will work as-is with the Premium Solver Platform.
You'll be able to solve nonlinear optimization problems twenty times larger than the standard Excel Solver -- up to 4,000 variables and 4,000 constraints. To handle such large problems, the Large-Scale GRG Solver Engine uses sparse matrix storage methods and advanced methods for maintaining numerical stability.
Sparse matrix storage methods take advantage of sparsity in large models, where subsets of the constraints typically depend on only a small subset of the variables. For example, the Jacobian matrix (the matrix of partial derivatives of the objective and constraints with respect to the variables) for a problem with 2,000 variables and 2,000 constraints would take about 32 megabytes for matrix storage using dense storage methods, but if this problem has the sparsity typical of larger models, it would take as little as 1 to 1.5 megabytes using the sparse storage methods in the Large-Scale GRG Solver.
Large nonlinear models require hundreds of thousands to millions of floating-point arithmetic calculations. Because of the finite precision inherent in computer arithmetic, small numerical errors occur in these calculations. Nonlinear models are particularly susceptible to the cumulative effect of these errors, which can lead to a numerically unstable or ill-conditioned matrix representation of the problem.
The Large-Scale GRG Solver uses several advanced methods to deal with numerical stability, including methods for estimating the conditioning of the Jacobian and Hessian matrix, special methods for dealing with degeneracy, and advanced methods for selecting changes in the basis (which represents the currently binding constraints in the problem). Thanks to these methods, the Large-Scale GRG Solver can find very good or optimal solutions to problems that could not be solved at all with more primitive nonlinear optimizers.
The GRG nonlinear Solver in the standard Excel Solver -- like virtually all "classical" nonlinear optimizers -- is guaranteed only to find a locally optimal solution. Imagine a graph of the objective function with "hills" and "valleys:" The GRG Solver will typically find the peak of a hill near the starting point you specified (if maximizing), but it may not find an even higher peak on another hill that is far from your starting point. In some problems this is sufficient, but in other cases you may want or need to find a globally optimal solution.
The Large-Scale GRG Solver includes "multistart" or "clustering" methods for global optimization. It can be automatically run many times from judiciously chosen starting points, and the best solution found (the "highest peak" if maximizing) will be returned as the optimal solution. For some smooth nonlinear problems, multistart methods will converge in probability to the globally optimal solution. For other problems, they often yield very good solutions in an acceptable amount of time – and of course, they are far easier to use than a manual exploratory process. And you don't have to change your model at all to take advantage of these new global optimization capabilities!
The Large-Scale GRG Solver offers several new options and tolerances to improve speed and give you better control of the solution process. The Recognize Linear Vars option allows the Large-Scale GRG Solver to take advantage of linearly occurring variables (often found in a large problem that is nonlinear overall), to save time in computing the Jacobian matrix. The Bypass Solver Reports option allows you to save time on runs when you don't need to produce report information. Other new options and tolerances control the multistart methods for global optimization.
If you've ever had problems with trial solutions that included small violations of the bounds you place on variables (sometimes generated by the standard GRG Solver), you'll appreciate the ability to control this behavior with the Relax Bounds on Variables option in the Large-Scale GRG Solver.
If you've found it difficult to enter your decision variable cells in the Solver because they are scattered around your worksheet, the Large-Scale GRG Solver Engine and the Premium Solver Platform are for you! They offer a variable cell list box entry alternative that allows you to enter an unlimited number of cell ranges. If you're tired of selecting Tools Solver... from the menu bar over and over as you solve various problems, you'll appreciate the "Return to Solver Parameters" check box in the Solver Results dialog, which lets you "stay within the Solver dialogs" until you need to work directly on your spreadsheet. And if your problem includes integer constraints, you can obtain a quick solution of the "relaxation" (temporarily ignoring the integer constraints) without having to delete these constraints and then re-enter them later.
Have you ever wondered about the size of the problem you have defined, and whether it is getting close to the size limits of the Solver? With the Large-Scale GRG Solver Engine and the Premium Solver Platform, you can check the size of your problem and the corresponding size limits in the Options dialogs at any time.
Have you ever wondered just how long "Setting Up Problem..." would take? If so, you'll appreciate the new progress reporting in the Large-Scale GRG Solver Engine V3.5, which shows an estimated "nn% Done" for both problem setup and generation of the Limits and Feasibility Reports, which can be time-consuming for larger models.
You'll appreciate the comprehensive 60-page User Guide that comes with the Large-Scale GRG Solver Engine, and supplements the 150-page User Guide that comes with the Premium Solver Platform. It provides complete information on Large-Scale GRG Solver options, the various messages which the Large-Scale GRG Solver can return, how to diagnose problems, and how to control the Large-Scale GRG Solver from Visual Basic in Excel.
The Large-Scale GRG Solver Engine V3.5 solves smooth nonlinear optimization problems of up to 4,000 variables and 4,000 constraints, and is priced at $995. It includes 90 days of knowledgeable technical support and automatic upgrades to new versions. You can order today with a 30-day money-back guarantee. Discounted prices are available for qualified academic users -- contact Frontline Systems for more information.
If you own another "Solver" product for Microsoft Excel, email or call us for the good news on competitive upgrade pricing! CONTACT US TODAY: Send us email, call us at (775) 831-0300, or send a fax to (775) 831-0314.
