The analysis of aftershocks is more and more recognized by the seismol
ogical community as being an essential tool to improve the understandi
ng of the earthquake faulting process, as well as of the various facto
rs contributing to the generation of earthquake damage. In this contex
t, the success or failure of aftershock monitoring programs depends to
a large degree on the performance of the mobile seismic network deplo
yed. In comparison with pure location networks, for aftershock monitor
ing networks the situation is complicated by the fact that they have t
o perform well for additional tasks such as determination of focal mec
hanisms, source-parameter estimation, tomographic experiments etc. We
have developed a method to design optimum networks for aftershock reco
rdings based on the technique of simulated annealing. If aftershock da
ta are not yet available for the design of a mobile network, they are
simulated for the optimization process from the mainshock-source param
eters. Different objective functions to quantify network performance f
or each intended task are used. For the location problem it is based o
n the average confidence volume of hypocentral locations while for the
focal mechanism determination it is derived from the distribution of
rays on the focal sphere. For the tomography problem, the resolution m
atrix of the inverse problem is used. Furthermore, for each station in
dividual noise levels can be assumed. Optimization for multiple tasks
is performed in a sequential way with highest priority given to the lo
cation problem. The method has been applied to the Roermond earthquake
of 1992 April 13. Optimum subsets of the existing ad hoc multinationa
l temporary network for aftershock recordings of this event have been
evaluated.