SLIC-AODV: a semi-local influence-based centrality-aware enhancement of AODV for mobile ad hoc networks

Mobile ad hoc networks (MANETs) require routing protocols that can sustain reliable packet delivery under mobility, contention, and dynamically changing topology. Although Ad hoc On-Demand Distance Vector (AODV) routing remains a widely used reactive protocol, its route selection is primarily governed by sequence-number freshness, hop count, and route-reply arrival order. As a result, AODV may select routes through structurally weak relay nodes when multiple candidate paths are available. This paper proposes SLIC-AODV, a Semi-Local Influence-based Centrality enhancement of AODV that integrates local/semi-local structural awareness into reactive route discovery. Each node computes a bounded routing-grade centrality score using influential spreading capability, semi-local influence, and local average shortest-path disruption. During RREQ forwarding, SLIC-AODV propagates a compact running-average path-centrality score, and the destination evaluates candidate routes using a composite \(F_1\) score combining hop-count efficiency and path centrality. A hop-aware route-upgrade rule prevents longer paths from being selected solely due to centrality. The protocol is implemented and evaluated in NS-3.41 over dense, sparse, mobility-varying, proactive-baseline, control-load, and ablation scenarios. Results show that SLIC-AODV improves AODV packet delivery ratio (PDR) in all dense scenarios, with the largest gain at \(N=50\) , where PDR increases from \(62.99\%\) to \(82.83\%\) . In moderately sparse scenarios, SLIC-AODV achieves significant PDR gains of \(3.86\) and \(6.76\) percentage points at \(N=30\) and \(N=40\) , respectively. It also reduces FlowMonitor-estimated normalized routing load from \(1.1085\) to \(0.1650\) in the dense \(N=50\) case, while introducing only marginal hop-count changes. Mobility experiments show AODV-like robustness, and ablation confirms that centrality contributes an additional \(9.97\) percentage-point PDR gain over the hop-only variant. These results indicate that SLIC-AODV is most effective in dense and moderately sparse MANETs where useful route diversity exists.