To mitigate the voltage violation and overloading caused by the integration of Distributed Generation (DG) and Electric Vehicle Charging (EVC) loads, a feasible solution is to convert some AC medium voltage distribution lines to DC ones and formulate an AC/DC hybrid distribution network, based on which flexible power shifting between lines can be achieved. However, the complementarity of power flows amongst AC/DC lines is hardly described and embedded in existing configuration optimization methods in which only the average loading or the variations of loading for a certain period is considered. Solution space will be geometrically increased if operation scenarios are fully simulated for configuration caused by the strong uncertainties of DGs and EVC loads. This paper proposes an analytic quantifiable method to describe the power complementarity between AC/DC lines. The Gaussian mixture model is used to depict the temporal correlation and uncertainties. The complementarity is calculated by the conditional probability of multi-dimensional joint probability density function of AC/DC hybrid lines. Accordingly, an optimal configuration method of hybrid AC/DC distribution network is proposed considering the complementarity of power flows on AC/DC lines, in which the size and location of Voltage Source Converters (VSCs) are optimized. Simulations studies are performed to verify the proposed method.