Bandwidth, Profile and Wavefront Optimization Using PSO, CBO, ECBO and TWO Algorithms

In this paper, three recently developed metaheuristic optimization algorithms, known as colliding bodies optimization (CBO), enhanced colliding bodies optimization (ECBO) and tug of war optimization (TWO), are used for optimum nodal ordering to reduce bandwidth, profile and wavefront of sparse matrices. The CBO is a simple optimization method inspired by a collision between two objects in one dimension. Each agent is modeled as a body which has a specified mass and velocity. A collision occurs between pairs of bodies, and the new positions of the colliding bodies are updated based on the collision laws. The ECBO employs memory to save some best-so-far position to improve the performance of the CBO without increasing the computational effort. This algorithm uses a mechanism to escape from local optima. The recently developed algorithm TWO is a multi-agent meta-heuristic algorithm, which considers each candidate solution as a team engaged in a series of tug of war competitions. The bandwidth, profile and wavefront of some graph matrices, which have equivalent pattern to structural matrices, are minimized utilizing these methods. Comparison of the achieved results with those of some existing approaches shows the robustness of these three new metaheuristic algorithms for bandwidth, profile and wavefront optimization.