چارچوب مفهومی برای مدل سازی مبتنی بر عامل خدمات لجستیک

This paper presents an agent-based microsimulation framework that represents the diversity of roles and functions of actors in the freight system, how they interact through markets and how interactions between actors are established in markets through contracts. The framework provides sensitivity to technology trends, business trends, and policy scenarios. Logistics costs, outsourcing of logistics services to third party logistics firms, growth or retraction of various industry sectors, and the impact of new supply channels are explicitly represented. Data sources available in Toronto, Canada and new data collection efforts required for model estimation are described.

Over the past few decades, freight transportation modelling has been approached from a variety of perspectives. Supply chain management and operations research approaches have been developed with the purpose of improving the efficiency of the business operations for single firms or groups of firms (see Meixell and Gargeya, 2005 and Crainic and Laporte, 1997 for good reviews). Models for public sector decision making have been developed in order to assess infrastructure improvements, land use alternatives, and other transportation policies such as road pricing (see TRB (2008) for a recent review of the current state of practice and future directions). Conventional approaches include three, or four stage modelling approaches (e.g. Pendyala et al., 2000 and Cambridge Systematics, 1996) and commodity based approaches (e.g. Cambridge Systematics, 1998 and Fischer et al., 2000). These approaches have, to a large degree, been modelled upon methods developed for passenger travel demand forecasting, even though there are fundamental differences between passenger travel and freight travel. More recently, a small number of hybrid models have been developed to attempt to incorporate behavioural elements of supply chain management and logistics chains into public sector decision making models (Hunt and Stefan, 2007, Fischer et al., 2005, Boerkamps et al., 2000, Wisetjindawat et al., 2007, de Jong and Ben-Akiva, 2007, Wang and Holguin-Veras, 2008, Yin et al., 2005, Tavasszy et al., 1998, Liedtke, 2006 and Bovenkerk, 2005). Hybrid models can recognize that: First, there are diverse actors involved in the production and distribution of goods, none of which may have full control or even knowledge of all decisions made throughout the supply chain. A single actor within the supply chain may be specialized in one component of the supply chain (e.g. a small carrier may only provide direct transportation services), may provide a suite of logistics services (e.g. consolidation, warehousing, transportation, distribution, inventory management), or may play roles in both production and distribution (e.g. a manufacturing firm with a private delivery fleet). Second, the interactions between firms are diverse. Successful supply chains increasingly involve long-term alliances between suppliers, manufacturers, retailers, carriers and 3rd party logistics firms. Other companies operate through spot markets for transportation services. The prices and the level of service vary depending on the type of relationship that is maintained between these business establishments. Third, business models are changing over time, including shifts in business operation towards lower levels of inventory, just in time delivery, and a business environment that is increasingly driven by customer orders (pull logistics). For many commodities order sizes are decreasing to accommodate customer demands for fast delivery. Globalization of trade has also increased, driven by reductions in the cost of transportation and geographical differences in resources, wage rates and production costs. Understanding and representing the roles that each actor plays in the freight transportation system, the interactions between those actors, and changes in the actors and their interactions over time are of fundamental importance in the development of more behavioural public sector models of the freight system. This paper presents an agent-based microsimulation framework that explicitly represents the diversity of roles and functions that business establishments play, how they interact through markets and how both long and short term interactions between business establishments are established in the market through contracts. The paper provides a framework only, and no specific application of the framework is presented. The framework is developed with the intention of providing a consistent modelling philosophy and a consistent degree of granularity with the ILUTE (Integrated Land Use Transportation Environment) modelling framework developed for the Toronto Area (Miller and Salvini, 2001 and Salvini and Miller, 2005). The ILUTE model is an agent-based microsimulation model of passenger activities and travel, which represents individual households, families, persons, and jobs, dwellings, etc. Designing compatible modelling systems from the outset will facilitate the eventual development of a fully integrated urban model of both passenger and freight activities and transportation (i.e. jobs modelled in the passenger system are consistent with labour requirements in the freight production system, commercial vehicle travel times are consistent with passenger vehicle flows, shopping trips are consistent with retail business establishment shipments, etc.). This paper is organized as follows. First, a literature review of other attempts to model the freight logistics system is provided. The conceptual framework is then presented, beginning with a definition of the actors (decision-makers), their attributes, functions and relationships. Following that, the other key model components are defined/formulated, including contracts, shipments, markets and logistics chains. The model’s temporal resolution is then presented with consideration given to the variety of decisions that are made over various time frames. The paper provides a plan for implementation, including potential data sources for estimation, and then concludes with a discussion of the potential for this modelling framework to address specific policy questions.

The key contributions of the conceptual framework provided in this paper to the body of literature on freight microsimulation modelling (described in Section 2) are as follows: (a) this framework is more comprehensive in nature, providing a consistent and systematic framework for representing business decisions ranging from fundamental long term decisions to short term operational decisions; (b) the framework outlines with mathematical notation the various roles that businesses play in their supply chains and in the economy as a whole, as producers, consumers and transporters of goods and services; (c) the framework’s explicit representation of contracts is a more realistic and explicit representation of commodity and service prices within markets than is present in other models. Translating this advanced conceptual representation into an operational model will be challenging, but rewarding. Application of the proposed conceptual framework would provide sensitivity to a variety of technology trends, business trends, and policy scenarios that more conventional approaches cannot do to the same extent. Specifically, • Cost of logistics services are explicitly represented, not only in the decisions associated with logistics chains, but also in the formation of longer shipper/carrier contracts logistics services markets. These costs are of key policy significance given recent diesel fuel price volatility, proposals for carbon taxes, road pricing initiatives, and costs associated with delays due to increasing congestion. Because activities of each firm are traced through the simulation system, it would be possible to assess differential impacts these trends/policies on business establishments from different industry sectors, different sizes, and different base locations. Microsimulation also implies that different values of time can be assigned to different logistics service providers carrying commodities with different values. • Trends toward the outsourcing of logistics services to third party logistics firms can be represented. Furthermore, long-term contract (i.e. shipper–carrier alliances) may have impacts on the efficiencies of logistics services, because of the economies of scale and better planning that can be achieved. • Because the model is commodity based, with input–output style sensitivities, it can assess the “trickle through” effects of the growth or retraction of various industry sectors, plant openings or closures. These effects would be spatially distributed. • The impact of new supply channels (e.g. e-commerce) can be modelled provided that logistics costs can be represented. • Public investment in infrastructure at intermodal terminals, exclusive truckways, etc. will differentially impact individual business establishments that operate or utilize different logistics chains. It will be possible in the microsimulation approach to show who the winners and losers are, and the extent to which such investment could be an “unfair” subsidy to one industry over another.