A two-stage deterministic algorithm for economical reinforcement design of MSE walls

Abstract

This paper presents a two-stage, deterministic optimization procedure for designing economical reinforcement layouts in Mechanically Stabilized Earth (MSE) walls. The method builds on a conventional analysis engine that performs internal stability checks (e.g., per AASHTO-type codes) and global stability using Limit Equilibrium Methods (LEM). In Stage 1, reinforcement spacing is optimized while length, tensile strength, and connection strength are held at defined upper bounds. By fixing these parameters to sufficiently large values, they are prevented from acting as bottlenecks to a stable solution. In Stage 2, with spacing fixed, tensile strength, connection strength, and reinforcement length are reduced using a batched 1-D bisection scheme that exploits approximate monotonicity and local separability of the stability response. Each variable is driven toward its minimum value consistent with the target factor of safety, yielding material savings without relaxing code-based safety criteria. Example applications demonstrate how the approach can refine or validate conventional MSE wall designs in a transparent and computationally efficient manner.