Abstract— Implementing and deploying a complete multi- agent system in whatever domain, especially those that are inherently complex and dynamic, is undoubtedly a very hard and time-consuming task. This paper analyzes the adequacy of traditional approaches in the field of Agent-Oriented Software Engineering to create adequate multi-agent systems applicable to the transportation domain. Our findings suggested that, albeit some approaches are intended to be generic enough to represent a wide range of different domains, intelligent transportation systems and more specifically the whole complexity of future urban transport are not totally supported. We propose a novel methodology where the concept of services is considered as peer of agents, ambient and processes. Also services rise as prominent elements in the modeling phase. The approach, illustrated as a typical scenario in the transport domain, is instantiated, which serves to clarify the whole process and main concepts in our modeling methodology. I. INTRODUCTION N respect to the global concern with ecosystems, the urban transport is faced as a “villain” due to the fact it is one of the main contributors for CO 2 emission [1]. Additionally, developing countries are experiencing growth in deaths caused by vehicle accidents, for instance as in Brazil and in China [2]. Observing this scenario, it becomes evident the lack of efficiency and the need for a remodelation of the transportation systems. Now, it is possible to use all new resources, namely communication networks, artificial intelligence, and others, to improve the system efficiency and safety. This new concept has already been presented within the Intelligent Transportation Systems (ITS) scoped. What is the real scenario expected for the future? What are all the features and technologies that can help us to achieve these goals? Yet, it is still too hard to answer all these questions. Nevertheless we can identify some main features of tomorrow’s intelligent transportation. As transport systems are rather spread out, Future Urban Transportation (FUT) requires a distributed architecture and advanced communication technologies to have the above requirement fulfilled, including accuracy as an important aspect to reduce system failure situations as well. Urban transportation systems are considered to be complex because they involve a large number of entities that L.S. Passos and R.J.F. Rossetti are with the Department of Informatics Engineering and the Artificial Intelligence and Computer Science Lab (LIACC) Faculty of Engineering, University of Porto, e-mail: {pro09026, rossetti}@fe.up.pt J.Gabriel is with IDMEC Pólo FEUP and the Associated Lab in Energy, Transportation, and Aeronautics (LAETA) Faculty of Engineering, University of Porto, e-mail: jgabriel@fe.up.pt act autonomously, a large number of influencing factors, e.g. weather and ecosystem. All relations between entities and influencing factors are nonlinear, dynamic, and hard to model precisely. Thus, some experimentation problems emerge [3], which are extremely difficult to tackle, and many times infeasible, to experiment in the real field due to its complexity. From this discussion arose the concept of Artificial Transportation Systems (ATS), and Multi-agent systems (MAS) paradigm is undoubtedly the best metaphor to represent and overcome all presented issues in an ATS framework. As any software, implementing agents must have an engineering process behind which is generally called Agent- Oriented Software Engineering (AOSE). It aims to understand the features that an agent-based approach can bring to the deployment systems taking advantage of autonomy, heterogeneity, and dynamism. Several AOSE methodologies have been proposed over the years [4, 6, 7, 8]. Due to the number of methodologies, the approach called Situational Method Engineering (SME) [5] aims to support the reuse of existing ones and, at the same time, to enable customization of a specific methodology, on basis of specific scenarios. Regarding urban transportation and its complex system characteristics, we state that there is no AOSE methodology which perfectly models the transport scenario and fulfils its requirements with respect to dynamism and interaction complexity. In the organization point of view, the available notation is insufficient to represent complex organizations and relations. Thus, a new methodology needs to be developed to take advantage of combined strategies from various areas to overcome the presented issues and this is what our approach aims to achieve. It is proposed an agent-oriented modeling methodology focusing on complex scenarios, such as artificial transportation systems, due to the lack of appropriate tools for this niche of applications. And, to do so, we use the SME approach based on Gaia [6], relating Gaia to service-oriented perspectives for its flexibility and being user-centred. Another reason is that the business process paradigm is being used by analysts and designers to easily model complex behaviors of the environment in most MAS applications. To illustrate our approach, we instantiate a simple scenario and give guidelines that can be of help to the reader in future modeling tasks. Following this brief motivation, the remaining of this paper is organized as follows. In the next section, we present the paper main section where the methodology is discussed, An Agent Methodology for Processes, the Environment, and Services Lúcio S. Passos, Rosaldo J. F. Rossetti, Member, IEEE, Joaquim Gabriel, Member, IEEE I