By definition, all constraints that are not linear are nonlinear. Nonlinear expressions include relationships in which variables are squared, cubed, taken to powers other than one, or multiplied or divided by each other.  Models with nonlinear expressions are much more difficult to solve than linear models. Unlike linear models, nonlinear models may prevent LINGO from finding a solution, though one exists. Or LINGO may find a solution to a nonlinear model that appears to be the "best", even though a better one may exist. These results are obviously undesirable. For more on what you can do to help minimize the occurrence of these undesirable results, see the Guidelines for Nonlinear Modeling section below.