Two commonly adopted expressions for the largest Lyapunov exponents of
linearized stochastic systems are reviewed. Their features are discus
sed in light of bifurcation analysis and one expression is selected fo
r evaluating the largest Lyapunov exponent of a linearized system. An
independent method, developed earlier by the authors, is also applied
to determine the bifurcation points of a van der Pol oscillator under
parametric random excitation. It is shown that the bifurcation points
obtained by the independent technique agree qualitatively and quantita
tively with those evaluated by using the largest Lyapunov exponent of
the linearized oscillator. (C) 1996 John Wiley & Sons, Inc.