Abstract：The structural robust optimization with uncertain parameters described by interval models is investigated. Comparing with the traditional optimization design, the non-probabilistic reliability index has been involved as an constraint condition in the robust optimization model. The relationship between the target of non-probabilistic reliability index and the fluctuation range of uncertain parameters is described by the geometric meaning of the reliability index. The non-probabilistic robust optimization design model is converted into a two-level optimization model. The minimization for non-linear performance function was carried out according to the fluctuant range of the non-probabilistic reliability index within the inner-lever optimization, so as to avoid the calculation of reliability index directly. For the linear performance function, the uncertain parameters could be describe as the target value of non-probabilistic reliability index with explicit expression, and therefore the two-level robust optimization is easily converted into a single-level deterministic optimization. The advantages of the proposed method lie in clear geometric characterization, computing convenience and high calculation efficiency. Numerical examples show the validity and efficiency of the proposed method.
2014, Vol. 35 
