Aggregation of Heterogeneous Firms in Mathematical Programming Models.

Serious aggregation errors may arise in mathematical programming sector models when aggregation is not properly done. Existing aggregation literature mostly focuses on exact aggregation of homogenous groups of firms. However, homogeneity rarely holds in real problem situations. Moreover, theoretical aggregation developments are generally based on the assumption of availability of full information about the firms to be aggregated. Such full information is hardly a plausible assumption when modeling the agricultural sector. This paper discusses an empirical aggregation approach based on decomposition theory, which may be particularly helpful while aggregating heterogenous firms under the absence of full information. Copyright 1989 by Oxford University Press.