A file assignment problem model for extended local area network environments
Krishna Rao Pattipati, Joel L. Wolf · 2002
A file assignment problem (FAP) designed specifically for file servers and work stations on an extended local area network (ELAN) is formulated and solved. Key properties of such an environment are modeled onto the FAP formulation. The FAP problem is NP-hard, and the approximate solution technique adopted uses Lagrangian (dual) relaxation. The dual FAP is solved by use of an accelerated subgradient method. The approach is efficient and also provides an estimate, called the approximate relative duality gap, of the quality of the solution. In all instances in which the method has been employed, the approximate relative duality gap is less than 1%. The algorithm is illustrated by several examples.>