For hazard assessment or countermeasure evaluation, estimating plausible flow paths of mass movement in landslide-prone areas plays a crucial role. Such estimation requires information on the geometry and depth of the failure surface, as well as the released volume. In this study, a simplex approach is proposed to approximate the failure surface using an Idealized Curved Surface (ICS; [1]), characterized by two distinct curves along the down-slope and cross-slope directions. The down-slope profile may be defined as a circular arc, a log-spiral curve or a profile computed by mechanical analysis (e.g., [2]), while a circular profile is adopted in the cross-slope direction. When circular profiles are used in both down-slope and cross-slope directions, the ICS can be uniquely defined by a reference ellipse and the maximum failure depth at its centre [3], with the reference ellipse determined from the boundary of the target area. Because the constructed ICS provides a smooth approximation of the rupture surface, a search procedure is employed to identify the best-fitted ICS subject to prescribed conditions, such as the delineated potential area, depths and released volume. The best-fitted ICS is obtained using a genetic algorithm (GA) through translation and rotation of the reference ellipses (GA-ER-ICS [4]), while parameters for constructing the corresponding log-spiral profile are provided by a trained neural network (NN) to improve computational efficiency. In addition, a 3D advanced scenario illustration platform (ANSIP) is developed to visualize computed failure surface and post-failure subsequent flow paths, which are simulated using a two-phase erodible model [5]. The integrated framework enables immersive exploration across various scenarios, supporting advanced real-time digital-twin applications for geohazard assessment. REFERENCES [1] Y.C. Tai, C.J. Ko, K.D. Li, Y.C. Wu, Y.C. Kuo, R.F. Chen, and C.W. Lin. An idealized landslide failure surface and its impacts on the traveling paths. Frontiers in Earth Science, 8:313, doi: 10.3389/feart.2020.00313, 2020. [2] K.C. Lin, Y.C. Tai, P.H. Lee, H.K. Wong, Y. Wang, Y.S. Lu, and J.S. Chen. An accelerated meshfree computational framework with machine learning classification for multi-phase modeling of landslide. Computers and Geotechnics, 190, 107756, 2026. [3] C.J. Ko, C.L. Wang, H.K. Wong, W.C. Lai, C.Y. Kuo, and Y.C. Tai. Landslide Scarp Assessments by Means of an Ellipse-Referenced Idealized Curved Surface.