Standard Excel Solver - Improving Slow Solution Times

Most Solver problems are computationally intensive -- and the time the Solver takes to solve a problem rises rapidly as the size of your spreadsheet and the number of decision variables and constraints increases. This page discusses methods you can use to reduce the time taken by the Solver.

Frontline's Premium Solver products can usually solve problems much faster than the standard Microsoft Excel Solver -- up to 100 times faster in some cases. Given the improved performance and low prices of new PC hardware, a faster computer is often a good investment.

Even so, the methods discussed here should help reduce the time taken to solve your problem with any set of hardware and software. The earlier in your model design process that you consider these factors, the better off you'll be when the model becomes larger (as it usually does):

• Linear Versus Nonlinear Models
• Eliminating Nonessential Calculations
• Adding or Tightening Constraints
• Methods for Nonlinear Problems
• Methods for Integer Problems
• Fast Problem Setup

The bottleneck in the solution process is nearly always recalculation time. While the Solver employs sophisticated, computationally intensive numerical methods internally, it still spends as much as 90% of its time waiting for Microsoft Excel to recalculate the worksheet. Therefore, any steps you can take to reduce recalculation time will favorably impact the Solver's solution time.