AppliedMath LexFlow: Deterministic Lexicographic Flow Allocation on Lossy Capacitated Trees
Annotatsiya
Exact, reproducible software for deterministic lexicographic flow allocation on lossy, capacitated rooted trees, and for the inverse rehabilitation-design problem posed on the same model. The allocation module implements the three-stage lexicographic model of the AppliedMath article Three-Stage Lexicographic Allocation on Directed Irrigation Trees with Conveyance Losses and Capacity Constraints: A Closed-Form Stage-1 Fairness Guarantee and Operator–Balance Equivalence: the closed-form Stage-1 bottleneck value, the loss-aware net-to-gross operator and its node–balance equivalent, the weighted Stage-2 and smoothing Stage-3 linear programs, the exact progressive-filling leximin selector, the invariant variation range and the price of fairness. It ships the five exact rational benchmarks, the 36-node, 16-period Gone Abat Jap controlled scenario, the deterministic scaling generator up to 500 users and 1022 edges, and the scripts that regenerate every published table and figure. The design module implements the inverse problem behind the conference paper Bottleneck-Targeted Rehabilitation Planning for Lossy Irrigation Canal Networks: An Exact Guarantee–Budget Curve: the closed-form minimum-cost widening plan and its budget function, the epsilon-bottleneck-set ascent for lining, and exact-rational relief sets and minimum-cost transversals with a laminarity certificate. Version 0.5.0 additionally provides bench/scale_timing.py, which measures the runtime and peak working memory of the closed-form Stage-1 evaluation against the equivalent sparse HiGHS linear program, and revises the published figures following peer review. All arithmetic on the small benchmarks is exact rational; inputs, outputs and the environment carry SHA-256 hashes.
Hali tarjima qilinmagan