Most of the conventional numerical techniques， such as the finite difference method， finite volume method， boundary element method and finite element method， are based on computational grids. Each grid point has a fixed number of predefined neighbours， and the connectivity between neighbounng points is used to define mathematical operators like derivation and integration. Often， Eulerian formulation is adopted， where the grid is fixed.Such an approach encounters computational complexities in modelling problems involving moving boundaries and large convections. These complexities can be avoided with the Lagrangian approach， where the computational grid moves with the material. However， the mesh can become tangled if the deformation of the material is large.
　　Rather than relying on the pre-defined grid connectivity， the meshfree methods rely on discrete particles to remove and thus completely avoid the mesh tanglement. Therefore， they are often referred to as particle methods in the literature. However， care must be taken that these particles are actually interpolation points for solving the partial differential equations.Only a small minority of meshfree techniques employ physical particles. Meshfree methods have been increasingly applied in computational hydraulics. In particular， the meshfree algorithms have demonstrated advantages in handling fluid/solid interactions and free water surfaces. However， they have also shown some limitations when being applied to other problems. As a relatively new numerical method， the meshfree technique is one of the hottest research topics in computational mechanics. In computational hydraulics， it has attracted more and more research interest. Over time， various types of meshfree algorithms have been developed， with different complexities and characteristics.