Asynchronous multi-agent systems have been widely used to optimize complex combinatorial problems. Their main strength lies in their ability to combine different heuristic algorithms and thus arrive at comparatively better solutions then their constituting algorithms on their own. A further strength is their flexibility in interlinked problems. This ability especially can be utilized in the semi-parametric modeling of applied spatial econometric problems. Semi-parametric modeling is required, if one wants to estimate nonlinear relationships between the dependent and the independent variables. The special case of spatial semi-parametric modeling requires determining the best configuration of independent variable vectors, number of spline-knots and their positions. Determining these poses an interlinked problem where the optimization of one part depends heavily upon the optimization of the others; therefore an asynchronous multi-agent team is used, to solve them simultaneously. This paper demonstrates that the proposed method can be used to model nonlinearity and furthermore it provides applied examples for real-life data sets.