Research Grants to Design Swarm-Intelligence Algorithms
Smart cities aim at enhancing the living standards of their citizens by making use of advanced technologies to innovate various sectors. Since these innovations often consist of optimising certain systems, or processes, the optimisation algorithms are widely used in applications relating to smart cities.
A prevalent optimisation application in smart cities can benefit various systems such as channel routing in networks and transportation systems. Routing can be represented as a Traveling Salesman Problem (TSP) which can be solved using optimisation algorithms. Another important aspect of smart cities is having good communication channels. Optical communication channels, especially optical spatial multiplexing systems are beneficial as they allow data to be transmitted quickly and safely. However, they rely on accurate channel estimation approaches to reduce signal degradation.
Artificial Neural Networks (ANNs) are popular algorithms used for channel estimation. Recently, swarm intelligence algorithms have been proven to be better training algorithms for ANNs as compared to conventional gradient-descent-based algorithms. The Dragonfly Algorithm (DA) is a swarm intelligence algorithm that is inspired by the behaviour of dragonflies in nature. The original DA algorithm is suitable for solving continuous optimisation problems such as the training of ANNs. Despite having good performance, the original DA has a low exploitation phase and its performance can be improved. Moreover, it cannot be applied to discrete optimisation problems such as TSP.
Principal investigator, Dr Muhammed Basheer Jasser and his team from several Malaysian and international universities are working on designing new swarm intelligence algorithms to address the problems of channel estimation and routing in smart cities. They have introduced an adapted discrete dragonfly algorithm and an enhanced swap sequence-based particle swarm optimisation algorithm for solving the TSP problem by addressing the routing problem. They also introduced an optimised continuous dragonfly algorithm to enhance the effectiveness of the original DA. Moreover, they have employed these algorithms in optimising channel estimation in optical spatial multiplexing systems and routing problems in smart cities.
The research has been funded by several research grants from Sunway University, namely Internal Grant Scheme and Kick Start Grant Scheme as well as the Publication Support Scheme. The outcomes from this research have been published in prestigious Q1 journals and presented at IEEE conferences, including the IEEE International Conference on Space-Air-Ground Computing held in Huizhou, China.