Transport policy analysis using multi-agent-based simulation
Transport policy analysis using multi-agent-based simulation
复制标题
使用基于多主体的模拟进行运输政策分析
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Linda Ramstedt
中科院分区：
文献类型：
--
作者：
Linda Ramstedt
本论文探讨了如何使用基于多主体的模拟来进行交通政策分析。交通政策通常被用作实现政府目标的手段，例如减少交通影响的环境目标。为了预测政策如何影响交通运输，公共当局经常使用模拟模型。对此类模型进行了结构化审查，重点关注重要的运输链特征。我们认为，为了正确预测运输政策对环境、经济和物流的实际影响，必须对运输链中的物流决策进行适当的建模。例如，此类决策涉及生产商和运输模式的选择、运输规划、生产和码头装卸。审查得出的结论是，目前用于交通政策分析的模型未能捕捉到其中许多特征。我们认为基于代理的模型有可能包含这些方面，因为它们能够明确地对运输链中的实际决策进行建模。我们在运输链中确定了一组通用角色，其中每个角色负责某些决策。基于多代理的模拟器 TAPAS 已经开发出来，其中这些角色被建模为代理。因此，捕获了运输链中的决策及其对运输政策应用的影响。这些决策导致物流运营的执行，进而对物流、经济和环境绩效产生影响。通过展示基于现实世界传输链的两个场景来说明 TAPAS 的用法。已经进行了引入不同类型的传输策略的场景的模拟实验。例如，通过将结果与类似研究进行比较以及输入参数的敏感性分析来分析模拟结果。为了促进模拟结果的验证和推广，我们建议利用典型的运输链和角色，例如产品类型和地理位置。描述了 TAPAS 可以支持的研究类型，并与通常使用传统模型进行的研究进行了比较。描述了与审查相关的运输政策，并分析了它们对运输链的潜在影响。讨论了 TAPAS 的可能用法并与不同类型的用户相关。例如，公共当局可以使用 TAPAS 来补充使用传统模型的研究。这可以通过包含更多逻辑方面来提高模拟结果的准确性。大公司是另一种类型的用户，例如，可以使用 TAPAS 来分析新的细分市场，例如新产品类型或新消费者，其中历史数据不可用 (A)。
This thesis explores how multi-agent-based simulation can be used for transport policy analysis. Transport policies are often used as a means to reach governmental goals, such as environmental targets to reduce the impact of transportation. To predict how transportation is influenced by policies, public authorities often make use of simulation models. A structured review of such models is made focussing on important transport chain characteristics. We argue that to properly predict the actual environmental, economic, and logistical effects of transport policies, the logistical decisions made in transport chains must be modelled appropriately. Such decisions, e.g., concern the choice of producer and traffic mode, planning of transportation, production, and terminal handling. The review concludes that models currently used for transport policy analysis fail to capture many of these characteristics. We argue that agent-based models have the potential to include these aspects since they are able to explicitly model the actual decision making in transport chains. We have identified a set of generic roles in transport chains where each role is responsible for certain decisions. A multi-agent-based simulator, TAPAS, has been developed in which these roles are modelled as agents. Thus, the decision making in transport chains and its influence by the application of transport policies are captured. The decisions lead to the execution of the logistical operations which in turn have consequences on the logistics, economic, and environmental performance. The usage of TAPAS is illustrated by presenting two scenarios based on realworld transport chains. Simulation experiments of the scenarios have been performed where different types of transport policies are introduced. The simulation results are analysed, e.g., by comparing the results to similar studies and by sensitivity analysis of input parameters. To facilitate the validation and generalisation of simulation results we suggest making use of typical transport chains and roles characterised by, e.g., product type and geographical locations. The type of studies that TAPAS can support are described and compared to studies typically made with traditional models. Transport policies which are relevant to examine are described and their potential influence on transport chains are analysed. The possible usage of TAPAS is discussed and related to different types of users. Public authorities can, e.g., use TAPAS to complement studies using traditional models. This can improve the accuracy of the simulation results by the inclusion of more logistical aspects. Large companies are another type of user which, e.g., can use TAPAS to analyse new market segments, such as new product types or new consumers, where historical data is not available (A).