The impact of distribution system characteristics on computational tractability

Mathematical programming models that seek optimal design and operational plans for distribution systems can be computationally intractable. This paper examines the extent to which distribution configuration and demand characteristics affect the ease of obtaining an optimal solution. Problem characteristics, which are reflected in a model by the parameter values, can render the model for one distribution scenario to be computationally intractable and that for another to yield an optimal solution easily. We introduce echelon-flow-based valid inequalities and use them to explicate the extent to which problem characteristics impact computational tractability.