AbstractA transshipment model was used to derive the existing and projected optimal structure of the US soybean processing industry. Analyses of model results for various time periods indicate that the potential exists for significant cost reductions on a national basis, if processing plants could be more optimally sized and located. It was also found that present processing capacity was in excess of present needs in most regions of the US; however, by the year 2000, almost all regions would need more plants or greater capacities than currently exist.