As traffic conditions constantly change, adaptive optimisation has proven to be an effective method for adapting traffic signal control systems accordingly. The utilisation of heuristic algorithms in directing traffic light control strategies, both in fixed time and real-time, has shown significant results. In order to improve the optimisation approach's ability to cope with modern traffic scenarios and synchronise with them, we propose a novel self-adapting algorithm to further enhance their capabilities. This work integrates particle swarm optimisation and ant colony optimisation with the novel self-adaptive approach, which enhances the selection of the most appropriate traffic cycle length to reduce traffic congestion based on real-world traffic conditions. The numerical experiments conducted on two traffic scenarios, peak hour and non-peak hour, show that our approach outperforms existing approaches by reducing travel time, traffic congestion, queue length, and pedestrian flow by 34%, 44%, 39%, and 11%, respectively. These results imply that our method can be implemented in real-world scenarios for sophisticated traffic light management.