设{X,Xn,n≥1}是独立的或φ-混合的或ρ-混合的正的平衡随机变量序列,或{X,Xn,n≥1}是正的随机变量序列使得{Xn-EX,n≥1}是平稳遍历的鞅差序列,记Sn=∑j=1^nXj,n≥1.该文在条件EX=μ>0及0<Var(X)<∞下,证明了部分和的乘积Пj=1^nSj/n!μ^n在合适的正则化因子下的某种重对数律.
Let {X, Xn, n ≥ 1} be a stationary stochastic sequence of independent, or φ-mixing, or p-mixing positive random variables, or {X, Xn, n≥ 1} be a positive random variable sequence such that {Xn - EX, n ≥ 1} is a stationary ergodic martingale differences, and set Sn = ∑j=1^n Xj j=l for n ≥ 1. This paper proves certain law of the iterated logarithm for properly normalized products of the partial sums,Пj=1^nSj/n!μ^n when EX = μ 〉 0 and 0 〈 Var(X) 〈 ∞.