Firm Synthesis and Asset Modeling

Firm Synthesis

The firm synthesizer enumerates the firms, and their member establishments and uses these datasets to create distributions of different firms’ attributes. These attributes are then simulated in an iterative procedure till the disaggregate totals for each attribute are matched. The main attributes synthesized are the employment, revenue, member establishments, location, and North American Industry Classification System (NAICS) codes. The framework synthesizes freight-intensive sector (FIS) establishments, i.e., establishments that produce physical goods as their primary products. The synthesis outputs are establishments and their parent firms belonging to 50 different 3-digits NAICS codes. Finally, FAF5 and CFS data are used to estimate the freight generation, i.e., tonnage produced and consumed by each establishment, as a rate per employment. POLARIS freight ABM framework mainly uses the following datasets for synthesizing firms: County and Zip Codes Business Patterns data, Statistics of U.S. Businesses data, Freight Analysis Framework 5 (FAF5) data, Commodity Flow Survey (CFS) data, Input-Output (IO) accounts data, and propreityr data from Dun & Bradstreet, FleetSeek, and CoStar.

Firm Asset Modelling

Freight ABMs frameworks for transportation operations rarely incorporate freight-related strategic asset decisions: fleet and distribution center (DC) ownership. These attributes are important for modeling freight transportation behavior in ABMs. We developed behavioral models that jointly predict fleet ownership and DC control for the U.S. FIS firms. A seemingly unrelated Tobit (SURT) regression is obtained under a Bayesian approach that allows quantifying the coefficients’ variability. Model results indicate that firms with higher revenue have an increased propensity to own fleet and (or) DC and prefer larger fleets and more DC space. Furthermore, Transportation firms generally have more heavy-duty trucks (HDTs) and much fewer medium-duty trucks (MDTs), while firms in all other sectors strongly prefer MDTs. In POLARIS, these models are used to estimate the number of HDTs, MDTs, and DCs areas owned by each firm. Using the results of these models along with locations and land use information, DCs are allocated to firms leveraging on sample data obtained from the CoStar dataset.

Important

Please refer to the the following papers for more details:

Stinson, M., and A. (Kouros) Mohammadian. Introducing CRISTAL: A Model of Collaborative, Informed, Strategic Trade Agents with Logistics. Transportation Research Interdisciplinary Perspectives, Vol. 13, 2022, p. 100539. https://doi.org/10.1016/j.trip.2022.100539.

Zuniga-Garcia, N., A. Ismael, and M. Stinson. A Freight Asset Choice Model for Agent-Based Simulation Models. Procedia Computer Science, Vol. 220, 2023, pp. 704–709. https://doi.org/10.1016/j.procs.2023.03.092