Multi-node multi-agent cloud simulation: Approximating synchronisation

Traffic engineering is a key in effective utilisation of the road network infrastructure. Simulation assists traffic engineers making informed decisions on how to operate and direct traffic within the road networks. These simulations are complex, generate big data and require high-powered computers, which can process information faster than real time, to ensure the results can be used to affect traffic. Cloud computing, a relatively new technology paradigm, can meet the essential requirements, such as scalability, interoperability, availability and high-end performance. In this paper, a novel approach to a synchronisation strategy of large-scale complex simulations is proposed. This approach builds upon advancements achieved in distributed computing. The new synchronisation strategy is designed to allow different granularities of synchronisation accuracy. Through this strategy, synchronisation overhead is reduced, thus allowing the computing bandwidth to be applied to simulation performance increases as a result of the trade off between synchronisation accuracy and performance.