A modified differential evolution-based combined routing and sleep scheduling scheme for lifetime maximization of wireless sensor networks

In recent years, wireless sensor networks (WSNs) have transitioned from being objects of academic research interest to a technology that is frequently employed in real-life applications and rapidly being commercialized. Nowadays the topic of lifetime maximization of WSNs has attracted a lot of research interest owing to the rapid growth and usage of such networks. Research in this field has two main directions into it. The first school of researchers works on energy efficient routing that balances traffic load across the network according to energy-related metrics, while the second school of researchers takes up the idea of sleep scheduling that reduces energy cost due to idle listening by providing periodic sleep cycles for sensor nodes. As energy efficiency is a very critical consideration in the design of low-cost sensor networks that typically have fairly low node battery lifetime, this raises the need for providing periodic sleep cycles for the radios in the sensor nodes. Until now, these two fields have remained more or less disjoint leading to designs where to optimize one component, the other one must be pre-assumed. This in turn leads to many practical difficulties. To circumvent such difficulties in the performance of sensor networks, instead of separately solving the problem of energy efficient routing and sleep scheduling for lifetime maximization, we propose a single optimization framework, where both the components get optimized simultaneously to provide a better network lifetime for practical WSN. The framework amounts to solving a constrained non-convex optimization problem by using the evolutionary computing approach, based on one of the most powerful real-parameter optimizers of current interest, called Differential Evolution (DE). We propose a DE variant called modified semi-adaptive DE (MSeDE) to solve this optimization problem. The results have been compared with two state-of-the-art and widely used variants of DE, namely JADE and SaDE, along with one improved variant of the Particle Swarm Optimization (PSO) algorithm, called comprehensive learning PSO (CLPSO). Moreover, we have compared the performance of MSeDE with a well-known constrained optimizer, called $$\varepsilon $$ε-constrained DE with an archive and gradient-based mutation that ranked first in the competition on real-parameter constrained optimization, held under the 2010 IEEE Congress on Evolutionary Computation (CEC). Again to demonstrate the effectiveness of the optimization framework under consideration, we have included results obtained with a separate routing and sleep scheduling method in our comparative study. Our simulation results indicate that in all test cases, MSeDE can outperform the competitor algorithms by a good margin.