Nowadays applications are strongly characterized by their complexity. They are usually composed by heterogeneous and distributed entities which must cooperate and coordinate in an ``intelligent'' way to exchange and share knowledge.
The agent-oriented paradigm [22, 16] is an emerging technology that faces the problem of modeling such kinds of applications. It is suitable for modeling entities (agents) which communicate (social ability), monitor the environment and react to events which occur in it (reactivity), are able to take the initiative whenever the situation arises (proactivity) without the intervention of human beings or other agents (autonomy). Societies of agents are called Multi-Agent Systems (MAS). They take into account the distribution of the involved agents and the integration of heterogeneous software and data, fundamental aspects for the success of today's software systems.
Unfortunately, at this time there is no evidence of a well-established engineering approach for building MAS-based applications. Due to the inherent complexity of those applications the prototyping approach seems more promising than the classical waterfall model thanks to its greater flexibility.
The paper presents the state-of-the-art of CaseLP (Complex Application Specification Environment Based on Logic Programming), an experimental logic programming based prototyping environment for MAS. The development of CaseLP started two years ago as an applied research on MAS rapid prototyping by means of logic programming [12]. The applications modeled during this period of time helped in understanding how to extend the framework. New features have been progressively added to cope with the emerging necessities.
CaseLP has been profitably used to develop prototypes in very different areas. Two applications were related to transport and logistic problems. The former has been developed in collaboration with FS (Italian railways) to solve problems of freight train scheduling along the railway between La Spezia and Milano [4]. The latter involved the planning of goods transportation, and has been realized in co-operation with Elsag s.p.a., an international company which provides service automation. Another application concerned the retrieval of medical information contained in distributed databases [18]. In this case CaseLP has been successfully adopted for a reverse engineering process. Finally the combination of agent-oriented and constraint logic programming techniques have been applied to the management of distributed database transactions [13].
The paper is structured as follows. Section 2 introduces CaseLP agents. Section 3 describes the six-steps method that is used to realize a prototype. Sections 4, 5 and 6 focus on languages and tools that can be used in different stages of the prototype development. Specifically, Section 4 presents tool and languages used in the modeling phase of the prototype. Section 5 describes instruments used in the implementation phase. Section 6 is related to the execution phase. Section 7 concludes the paper outlining the related and future work.