Power-aware river formation dynamics algorithm advances energy-conscious mobile networking

A recent study published in the International Journal of Computer Networks and Applications (Vol. 11, No. 3, 2024) introduces the Power-Aware River Formation Dynamics-Based Routing Algorithm (PRFDA), a nature-inspired metaheuristic developed to enhance energy efficiency in Mobile Ad hoc Networks (MANETs). PRFDA reimagines data routing through the lens of natural water flow, where drops of water find optimal paths by eroding and depositing soil, a process now used to model intelligent and energy-sensitive packet forwarding in wireless networks.

MANETs are self-organizing systems composed of mobile, battery-powered nodes that must operate without centralized infrastructure. Their deployment in emergency relief, battlefield communication, rural education, and healthcare scenarios makes energy conservation not just a performance requirement but a survival necessity. PRFDA addresses this challenge by applying the River Formation Dynamics (RFD) model to routing, introducing a novel mechanism that prioritizes energy preservation, delay minimization, and efficient hop reduction.

Each node in the network is assigned a virtual altitude derived from a cost function that integrates three key metrics: the node’s remaining energy, its estimated transmission delay, and the number of hops required to reach the destination. Nodes with higher energy, lower delay, and fewer hops are considered more attractive and are thus assigned lower altitudes. In this metaphorical landscape, data packets are modeled as water drops, which naturally move toward nodes with lower altitude, just as rivers flow downhill.

A defining innovation of PRFDA is not merely the classification of neighboring nodes into categories based on their altitude gradients, but the way it assigns different erosion and probability computation strategies to each category. These computations directly incorporate energy, delay, and hop count, ensuring that each neighbor type is evaluated with an awareness of both its gradient and its practical cost. This context-sensitive behavior allows PRFDA to intelligently favor paths that optimize overall network performance and avoid routes that could lead to premature node depletion.

Routing begins when a source node launches multiple drops, each of which explores the network by probabilistically selecting the next hop. This selection is governed by a transition function that factors in altitude differences, historical soil levels, and the node’s category. Drops prioritize neighbors that lead downhill and represent energy-efficient routes. When such neighbors are unavailable, the algorithm allows drops to fall back on flat or uphill nodes, preserving network connectivity and avoiding routing failure.

As drops move through the network, they apply erosion to the links they traverse, adjusting soil levels based on the quality of the neighbor and the drop’s velocity, which itself is a function of energy and delay. The more desirable the path, the more soil is removed, lowering resistance and reinforcing that link's attractiveness. Simultaneously, sediment is deposited on overused paths to prevent them from being overexploited. This dual mechanism of erosion and sedimentation supports adaptive route formation while promoting load balancing and network resilience.

Each drop keeps track of its path, and loops are actively prevented by checking visited nodes. Drops that become trapped, unable to find eligible neighbors, are considered blocked and excluded from further processing, preventing energy waste. Once all active drops reach the destination or are blocked, their paths are evaluated. The best route is selected using a total cost function that combines cumulative delay, hop count, and the minimum energy of nodes along the path. This chosen route is then reinforced in the routing table, with corresponding updates to soil and altitude, ensuring the system learns and adapts over time.

PRFDA was tested using the NS-3 simulator, with results compared to four existing protocols: EMBO, RFD, AODV, and DSDV. Simulations under varying pause times and traffic loads demonstrated that PRFDA consistently achieved higher packet delivery ratios, lower end-to-end delays, better energy utilization, and longer network lifetimes. These results validate its effectiveness as a power-aware enhancement to conventional RFD-based models.

Future extensions proposed in the study include integrating congestion-aware queue management, adaptive traffic class handling, and hybridizing PRFDA with other AI-based models to improve responsiveness in dynamic topologies. By grounding its routing intelligence in the physics of water flow while embedding power-awareness at every step, PRFDA offers a practical, scalable, and scientifically grounded solution to the challenges of energy-conscious wireless networking.

About the author

Dr Augustina Dede Agor is a lecturer in the Department of Information Technology at the University of Professional Studies, Accra. She holds a PhD in Computer Science. Her research focuses on metaheuristics, computer networks, wireless communication, artificial intelligence, security, energy- aware routing, and optimization techniques for mobile and ad hoc networks.