Constraint equations are increasingly being used in interactive applications such as graphics, logical programming, and simulation that demand immediate feedback. To handle the performance requirements imposed by such systems constraint evaluators must use incremental satisfaction techniques. In this paper, we apply these techniques to noncircular, multilinear systems of equations. The constraint satisfaction process is divided into two phases--a planning phase that imposes a topological order on the equations and an execution phase that evaluates the equations. A planning algorithm is presented that incrementally updates this order each time the constraint system changes. This technique achieves significant performance improvements in large constraint systems since modifications generally perturb only a small portion of the topological order.