Cutting down the cost and the time necessary for generating numerical grids is an emerging requirement since parallel processing has enormously cut down the cost and time of physical analysis after a grid is generated. Besides, emerging applications in the biomedical, bioengineering and nanosciences arena demand new approaches to grid generation. These applications demand higher accuracy of surface representation and typically involve highly flexible/deformable surfaces and require a higher degree of automation. A new advancing front approach is proposed here that will address these emerging requirements on numerical grid generation. Moreover, the proposed approach can be used in other related areas such as in image processing for identifying surfaces of interest, in computational geometry for "surface repair", and, in computing fluid flows involving moving/deforming boundaries. The formulation presented addresses both structured and unstructured/hybrid grids. The proposed effort will result in a software product that will be developed and used for proof of concept using simple geometries with the help and involvement of graduate and undergraduate students, who will be thoroughly exposed to the practical use of modern software design principles as well as mathematical and numerical analyses. Scientific knowledge and experience accrued during the course of this investigation will be disseminated to the community through conference presentations, with the involvement of student presenters, and journal publications, as well as be made available over the Internet.