Adaptive node moving refnement in discrete least squares meshless method using charged system search

Discrete Least Squares Meshless (DLSM) method has been used for the solution of dierent problems ranging from solid to uid mechanics problems. In DLSM method the locations of discretization points are random. Therefore, the error of the initial solution is rather high. In this paper, an adaptive node moving renement in DLSM method is presented using the Charged System Search (CSS) for optimum analysis of elasticity problems. The CSS algorithm is eectively utilized to obtain suitable locations of the nodes. The CSS is a multi-agent optimization technique based on some principles of physics and mechanics. Each agent, called a Charged Particle (CP), is a sphere with uniform charge density that can attract other CPs by considering the tness of the CP. To demonstrate the eectiveness of the proposed method, some benchmark examples with available analytical solutions are used. The results show an excellent performance of the CSS for adaptive renement in meshless method.