Content area
In order to access the power of the tools provided by the Optimization Subroutine Library (OSL), an OSL-based application must pose problems in a form that OSL recognizes and interpret the solution that OSL provides. A systematic approach is presented for developing OSL applications that can be applied to most formulations of mathematical programming problems. From a symbolic representation of the problem, the application developer can generate code that will perform the formatting of data for OSL, solve the problem, and retrieve the solution. The emphasis is on seamlessly integrating OSL with the rest of the application by working with data directly, while the data may be stored elsewhere in the application. The approach can be applied to any linear programming problem that can be represented in the symbolic format described. The same techniques can be modified for writing applications for quadratic and linear mixed-integer programming.