Due to the size and number of interdependencies in a complex system, the resilience of such a system to disturbances is often questionable. In this paper, a case study, an industrial ecosystem in Kalundborg, Denmark, is modeled and investigated with respect to its ability to handle disturbances. Several disturbances are introduced and the robustness through time of production processes to these disturbances is measured and improved with an augmented form of the Robust Concept Exploration Method, in which response surfaces and a multiobjective problem solving tool are utilized. The standard RCEM is augmented to reduce the degree of approximation on which the solution ranges are based. This approach is significantly different from others in that behaviors over time, not instantaneous values, are improved. These behaviors over time are evaluated using multiple time pattern descriptors which characterize relevant features of the behaviors. The simulation model used in this paper is not emphasized as much as the processes used in measuring and improving the behavior over time of a complex system.