Since we could not know the error between the evaluated value v (k) ijq and the population objective value V (k) ijq (unknown), we therefore cannot take the confidence level as 1, that is, we cannot take the membership grade of v * (k) ijq as 1. Thus, it should be more reasonable to consider the membership grade of v * (k) ijq within the interval (λ, 1), 0 < λ < 1. In this study, we presented level (λ, 1) interval-value fuzzy numbers and compositional inference rule to tackle the aggregative risk rate in the fuzzy sense.