Network-Lifted Digital Twins for Logistics: Feasible Pack Sets, Residue Spectra, Annealed Allocation, and Additive Attribution
Abstract
A supply-chain digital twin is treated here as a clock-aligned replica of plants, docks, lanes, inventories, and policies. This paper treats that replica as the network lift of a vehicle energy-store problem: pack polytopes, channel integrity, command latency, and solar harvest already define a feasible set; leftover charge after a stint is the same residual object as leftover pallets after an assignment. We write the twin state, an observer, a three-coordinate resilience vector, a disruption-aware assignment Hamiltonian, a two-sided shift identity that moves delayed residue pulses without leaving the test space, and a Shapley head that attributes service movement to pack, channel, harvest, and disruption features. The literature is organized as one regularized program: size the store, keep the channel observable, read the residual spectrum, twin the flow, anneal the allocation, and attribute leftovers so that delayed residue remains interchangeable with on-time residue. Keywords: digital twin; supply-chain resilience; pack sizing; residual spectrum; ripple effect; quantum allocation; two-sided FKF transform; Shapley attribution; TreeSHAP; KernelSHAP