Summary: We consider a capacity allocation problem of a generalized Jackson network (a Jackson network with general interarrival time and service time distributions). The problem is to determine the service rate (or capacity) that minimizes the expected equilibrium customer delay subject to a linear budget constraint on the capacities. The problem is analyzed using a Brownian approximation of the generalized Jackson network. When the network data satisfy a certain condition, the resulting capacity allocation is a generalization of the classical square root capacity allocation for Jackson networks. A numerical example is provided that demonstrates the allocation’s effectiveness.