Research Paper: Computational efficiency enhancement of graph-based network enumeration algorithm (Jan. 2024 - May 2024)
Technologies and Skills: Graph Theory / Network Design, MATLAB, Algorithm Optimisation, Runtime Analysis, Matrix Structural Analysis, Parallel Processing, Suspension systems, Technical Report Writing
This project was my first scientific research paper, which presented the opportunity to contribute to a real-world problem. My research paper focussed on improving Passive Vibration Absorber (PVA) design. PVAs use elements such as springs and dampers to control the energy absorption in a given application. The paper identifies several applications for PVAs such as suspension systems in road cars and trains, as well as vibration suppression in buildings and bridges. In addition, a newly introduced element known as the inerter was considered, which provides a force proportional to the relative acceleration between its terminals.
The research paper was based on work conducted by my supervisor, Dynamics and Control professor Dr Jason Zheng Jiang of the University of Bristol. His research documents an enumeration algorithm which is able to explore every possible network configuration comprised of springs, dampers and inerters. However, this algorithm has exponential time complexity and thus runtimes exceed permissible limits with approximately 5-6 elements. Therefore, my research paper aimed to investigate the runtime complexity of the algorithm and attempt to reduce runtimes.
The objectives of this research paper were therefore to: 1. Review and implement the graph-based enumeration algorithm to investigate its runtime weaknesses. 2. Improve upon the algorithm by identifying computational methods to decrease runtimes. 3. Test the resulting runtime improvements. 4. Demonstrate and discuss the performance benefits resulting from increasing the specified number of elements.