ABSTRACT
Superstructure-based models in PI provide an alternative approach to PA. Such models rely on a direct representation of system components in a pre-defined network that contains all possible connections. During optimization, excess connections are removed, and the flow rates associated with those that remain are determined along with the optimal solution. This approach leads directly to an optimal network if a solution algorithm that guarantees global optimality for the problem class is used. In this chapter, problems solved earlier using PA are revisited and solved using superstructural models. Such models can be solved either in spreadsheets (Excel) or in optimization software (LINGO). The advantages and disadvantages of superstructural models compared to PA-based methodology are also discussed.
