Develops a framework for averaging functional differentialequations (FDEs) with two time scales. Averaging is performed on thefast time system, while slow time is `frozen.' This creates an averagedequation which is slowly time-varying, hence the terminology of partialaveraging. We show that solutions of the original FDE and itscorresponding partially averaged equation remain close on arbitrarilylong but finite time intervals. Next, assuming that the partiallyaveraged system has an exponentially stable equilibrium point and thatwe restrict our interest to initial conditions that lie in the domain ofexponential stability, the finite-time averaging results are extended toinfinite time. In the special case of pointwise delays, exponentialstability of the averaged system can be related to the frozen-timeeigenvalues of its linearization
展开▼