Aggregation in the generalized transportation problem

Abstract

Aggregation disaggregation is used to reduce the analysis of a large generalized transportation problem to a smaller one. Bounds for the actual difference between the aggregated objective and the original optimal value are used to quantify the error due to aggregation and estimate the quality of the aggregation. The bounds can be calculated either before optimization of the aggregated problem (a priori) or after (a posteriori). Both types of the bounds are derived and numerically compared. A computational experiment was designed to (a) study the correlation between the bounds and the actual error and (b) quantify the difference of the error bounds from the actual error. The experiment shows a significant correlation between some a priori bounds, the a posteriori bounds and the actual error. These preliminary results indicate that calculating the a priori error bound is a useful strategy to select the appropriate aggregation level, since the a priori bound varies in the same way that the actual error does. After the aggregated problem has been selected and optimized, the a posteriori bound provides a good quantitative measure for the error due to aggregation.