If you're looking for an Excel Solver upgrade, you've come to the right place.  Frontline Systems developed the Solver in Excel for Microsoft, and we're experts in optimization.  Our Premium Solver products are upward compatible from the Excel Solver, and easy to use.  You may have any of these problems:

Our Premium Solver products can definitely help with all four of these problems.  Size limits are the easiest -- Our products can handle up to millions of decision variables.  Solution time is often easy to improve -- our Solvers use faster algorithms and exploit multi-core processors -- but it may also depend on your model design.  Poor solutions can be improved, but this may call for changes in your model design.

Let's tackle the last issue first.  Do you know that you've built a linear, or smooth convex optimization model in Excel?  Click your answer in one of the links below:

Solving Larger Linear or Smooth Convex Models

This is straightforward.  With Premium Solver Pro, you can solve linear programming (LP) models up to 2,000 variables, and smooth convex nonlinear (NLP) models up to 500 variables.  With Premium Solver Platform, you can solve LP models up to 8,000 variables (maximum 2,000 integer variables) "out of the box," and with its plug-in Solver Engines, you can solve much larger models, with no effective limits.  Premium Solver is faster, and Premium Solver Platform is much faster than the standard Excel Solver.  Click here to learn more and download a free trial of our software.

Solving Larger Non-Convex or Non-Smooth Models

This is harder (The problem space takes on many dimensions!), but our Premium Solver products can help.  With Premium Solver Pro, you can solve non-convex nonlinear models with the GRG Solver and Multistart, and non-smooth models with the Evolutionary Solver, both with up to 400 decision variables.  Solution times for the Evolutionary Solver are often faster, but solution times with the GRG Solver may be similar to those in the standard Excel Solver.  With Premium Solver Platform, you can solve problems with up to 500 variables faster if you have a multi-core processor; and with its plug-in Solver Engines, you can solve much larger models at potentially much faster speeds.  Click here to learn more and register to download a free trial of our software.

I Don't Know if My Model is Linear or Smooth Convex

You need to know.  Our software can determine this for you.  Click here to learn more and register to download a free trial of our "super-product" Risk Solver Platform (RSP - a superset of Premium Solver Platform), and use it to diagnose and solve your model.  When it starts up, RSP runs in "Guided Mode," which displays a series of dialog boxes explaining what it finds.  Pay attention to what those dialogs say.

You may learn (i) that your model is linear or smooth convex, and can be solved readily, (ii) that your model is non-smooth but RSP was able to automatically transform it to a better form, or (iii) that your model is non-convex or non-smooth, and it's going to be difficult to solve unless you do some work to avoid use of certain functions (such as IF, CHOOSE and LOOKUP) in certain parts of your model.

I Don't Care if My Model is Linear or Smooth Convex

You need to care.  Let's use an analogy:  Imagine that the Excel Solver was a jet aircraft, but it wasn't fast enough for you.  You've heard that we offer supersonic aircraft, so you visit our showroom.  "I have two travel plans," you say.  "In my smaller test, I need to travel from New York to Beijing in 20 minutes.  In my full-scale test, I need to travel from New York to Alpha Centauri in one hour."  We reply, "Well, New York to Beijing in 20 minutes is tough (20,000 miles per hour), but we may be able to help you.  But New York to Alpha Centauri in one hour is impossible -- that's more than 4 light-years away."  If you tell us "I don't care how far it is, I just want you to get me there," you're ignoring the laws of physics, and we can't help you.

If you don't care whether functions like IF, CHOOSE and LOOKUP (that depend on decision variables) make your model non-convex and non-smooth, you're effectively ignoring the laws of mathematics.  We actually can help you with small models, but we can't help you with large ones.  For a non-convex model -- except in special cases -- solution time grows exponentially with model size.  When your model is large enough, this factor swamps all others -- faster processors, faster algorithms or solution methods, and the like can't keep up with exponential growth.  Please see the next topic about poor solutions, because this usually arises in the same situations -- and click here to learn about optimization problem types.

My Model Yields a Poorer Solution Than Expected

You need to answer the question posed above:  Do you know that you've built a linear, or smooth convex optimization model in Excel?  If the answer is "yes", you (or we) need to take a closer look at your model.  Usually we find that:

  • You have an error in your Excel formulas or constraint selections, and fixing this will yield much better solutions.
  • You have a poorly scaled model, and fixing this and/or applying an enhanced Solver Engine will yield better solutions.

If the answer is "no" or "I don't know," chances are rather good that you've built a non-convex or non-smooth model, and this is why you are getting a poor solution (possibly in combination with other errors or poor scaling).  If your model isn't too large, we can usually help you rework or simplify your model and get better solutions.

Click here to learn about optimization problem types.  As explained above, our software can automatically determine the type of model you've built.  Click here to learn more and register to download a free trial of our "super-product" Risk Solver Platform (RSP - a superset of Premium Solver Platform), and use it to diagnose and solve your model.

How Do I Explain the Benefits of This to Management?

We have a brief overview page for managers.  You can:

These are bottom line issues in almost any company.

How Can I Find Out if Your Software Will Work for Me?

Congratulations for reading this far -- you've started to gain the knowledge to be successful with optimization.  When you have the knowledge and tools to build an optimization model to allocate scarce resources, the payoff can be enormous.  Many of our users have saved their companies millions of dollars.

Register now! -- it's simple and free -- so you can:

  • Download a free 15-day Trial Version of any of our Excel Solver products
  • Download our Solver User Guides -- learn more about our Excel products, and how to build better optimization models
  • Access "protected" Tech Support pages and downloadable example models

Just enter your name and email address, and optionally your company and phone, pick the closest "User Type," and click the Register for Download button.  You can download immediately after registering, or return later anytime -- just login with your email address -- when you're ready to try out our Excel Solver upgrades.