This message appears when Solver recalculates your worksheet using a new set of values for the decision variable cells, and discovers an error value such as #VALUE!, #NUM!, #DIV/0! or #NAME? in the objective cell or one of the constraint cells. Inspecting the worksheet for error values like these will usually indicate the source of the problem. If you’ve entered formulas for the right hand sides of certain constraints, the error might have occurred in one of these formulas rather than in a cell on the worksheet. For this and other reasons, it’s better to use only constant numbers and cells containing constant numbers on the right hand sides of constraints.
If you see #VALUE!, #N/A or #NAME?, look for names or cell references to rows or columns that you have deleted. If you see #NUM! or #DIV/0!, look for unanticipated values of the decision variables which lead to arguments outside the domains of your functions – such as a negative value supplied to SQRT. You can often add constraints to avoid such domain errors; if you have trouble with a constraint such as $A$1 >= 0, try a constraint such as $A$1 >= 0.0001 instead.
The Evolutionary Solving method rarely, if ever, displays this message – since it maintains a population of candidate solutions and can always generate more candidates, it can simply discard trial solutions that result in error values. If you have a model that frequently yields error values for trial solutions generated by the GRG Nonlinear method, and you are unable to correct or avoid these error values by altering your formulas or by imposing additional constraints, you can still use the Evolutionary Solving method to find (or make progress towards) a “good” solution