Evolvable Multi-Agent Hierarchical Architecture for Contested Logistics Ecosystems
1 INTRODUCTION: FROM SUPPLY CHAINS TO CONTESTED LOGISTICS
Supply chains are end-to-end systems through which value flows from raw materials to end recipients, with logistics serving as the execution layer responsible for coordinating movement, storage, and resource allocation across time and space. When logistics systems fail to adapt, overall supply chain resilience is compromised. As supply chains evolve toward more distributed, data-driven, and collaborative models, supporting logistics architectures must evolve accordingly.
Global supply chains now operate under increasing volatility. Geopolitical instability, climate disruption, cyber threats, and interdependent markets have exposed the fragility of systems built on assumptions of stable infrastructure and reliable data flows. [ 1 ] As a result, many logistics systems struggle to perform effectively under uncertainty, disruption, and fragmented information.
In this context, contested logistics environments describe settings where systems must operate despite disruption or degradation of physical and digital resources, whether from infrastructure failure, data friction, or intentional adversarial action. Although originating in military doctrine, [ 2 ] these conditions extend beyond defence. Commercial, humanitarian, and critical infrastructure supply chains face analogous challenges. Contested logistics should therefore be understood as a general operating condition across modern supply chains.
This paper introduces an evolvable multi-agent hierarchical architecture, supported by an AI- Ready Data Intelligence Platform, to enable coordinated reasoning under both operational uncertainty and adversarial pressure. The architecture integrates autonomous agents, human decision-makers, and trusted data services across structured layers. Its domain-independent design offers a transferable blueprint for circular, smart, and adaptive supply chains. This work contributes by reframing contested logistics as a cross-domain challenge, presenting a coherent architectural approach that addresses both degradation and deception, and grounding it in a realworld modernisation programme.
2 RELATED WORK AND BACKGROUND
The challenges of managing complex logistics under uncertainty have attracted sustained research attention across supply chain management, multi-agent systems, and artificial intelligence. This section positions the proposed architecture within that literature.
2.1 SUPPLY CHAIN RESILIENCE
Christopher and Peck [ 1 ] established that resilience requires not only physical redundancy but information agility. Sani, Schaefer, and Milisavljevic Syed [ 2 ] extended this to military supply chains, arguing for pre-emptive resilience strategies. Both works highlight that resilience must be designed into systems from the outset, not retrofitted after disruption.
EDM Association – Journal of Innovation 61