1,721,185 research outputs found
Designing a Multi-Agent System for Integrated Protection in Agriculture
No full textIntegrated Protection (IP) in agriculture consists of a set of practices aimed at favoring the set up of a development model characterized by a reduced environmental impact. The application of IP practices in plant disease management by growers and agronomists requires both specialistic skills, historical data and information on chemicals and on low impact techniques for pest management. These sources of information and knowledge are distributed among different actors in the agriculture production system.
Recent approaches in developing decision support systems for agriculture, and more generally for environmental problems management, tend to adopt a "systemic" approach. That is to say a problem is considered for its dependencies to different skills and responsibilities, and the proposed applications aim to be integrated in larger information systems. So basically, two main dimensions of complexity have to be considered while analyzing the problem: the organizational dimension dealing with all the dependencies between the domain stakeholders, and the technical dimension concerning the study of natural plant protection techniques.
These considerations motivate our choice of using an agent-oriented methodology for software development in designing a multi-agent system at support of apple growers and technicians of the advisory service. The methodology, called Tropos, gives a central role to early requirements analysis and allow to derive system functional and non-functional requirements from a deep understanding of the domain stakeholders goals and of their dependencie
Software Architecting and Designing adopting an Agent-Oriented approach
No full textSoftware design is a complex problem-solving process which requires to evaluate several design options while pursuing the objective of adhering to general principles of good quality. For instance, designing open, robust and secure architectures for novel application areas, calls for evaluating and/or integrating different distributed system technologies, such as peer-to-peer and multi-agent systems. This paper proposes a discussion on software architecting and designing strategies adopting an agent-oriented approach. Several works face this issue proposing methodologies based mainly on the use of architectural styles. Here we consider general strategies useful in the design process, such as divide-and-conquer, and principles for a good design, such as cohesion and decoupling, and argue that, performing goal analysis, according to Tropos, an agentoriented methodology, provides methods for applying them. A case study in the domain of decision-support systems in agriculture is used to illustrat
Developing a Decision Support System for Integrated Production in Agriculture
No full textRecent approaches in building decision support systems (DSS) for agriculture, and more generally for environmental problems, tend to adopt a ``systemic'' approach. That is to say a problem is analyzed in terms of all the knowledge, the data and the responsibilities it depends on. So, the proposed applications aim to be integrated in larger information systems exploiting the fact that different organizations may manage information sources and resources that are relevant to problem solutions.
The paper focuses on design issues faced during the development of a DSS at use of technicians of the advisory service performing pest management according to an Integrated Production approach.
Designing this type of system requires to analyze basically, two main dimensions of complexity: the organizational dimension dealing with all the dependencies between the domain stakeholders, and the technical dimension concerning the study of natural plant protection techniques.
These considerations motivate the choice of an agent-oriented methodology for software development. The methodology, called Tropos, gives a central role to early requirements analysis and allows to derive system functional and non-functional requirements from a deep understanding of the domain stakeholders goals and of their dependencies.
Two components of the system have been implemented using web technologies and they are currently under evaluatio
PICO Project: Metodologia di valutazione del Prototipo Software
No full textLa validazione del sistema PICO si estrinseca nella valutazione delle due componenti software sviluppate in accordo con le specifiche descritte nel deliverable W3D1. Si tratta, in particolare del modulo basato su GIS, qui denominato GISP (Geographic Information Service Provider) e del modulo basato su tecniche di Machine Learnig, qui
denominato DBL (Disease Behavior Learner), che saranno valicati da parte
degli utenti finali del sistema stesso. La fase di valutazione del prototipo del sistema web based GISP, per la progettazione dell'impianto di confusione sessuale ad uso dei tecnici del Centro Assistenza Tecnica dell'Istituto Agrario San Michele all'Adige (CAT-IASM, ex ESAT) persegue due obiettivi illustrati nel seguito. Da un lato la verifica delle funzioni fornite dal sistema rispetto agli user requirements. L'obiettivo è quello di ricevere risposte del tipo: la progettazione è conforme, non è conforme (e di quanto si discosta) rispetto ai requisiti utente. Dall'altro la verifica della progettazione della GUI (Graphic User Interface) e sue funzionalità collegate, la raccolta di un'eventuale insieme di nuovi requisiti.
La procedura di validazione del prototipo della componente DBL ha lo scopo di valutare la significatività dei risultati dell'indagine sui dati riguardanti alcune fitopatie considerate nel progetto basata sull'uso di tecniche di Data Minin
PICO Project: Analisi requisiti addizionali per la messa in rete del sistema
No full textL'analisi dei requisiti per la messa in rete del prototipo sviluppato nell'ambito del progetto PICO ha lo scopo ultimo di comprendere quali siano gli aspetti di cui tenere conto nel passare dalla fase prototipo alla fase di sistema effettivamente in uso. Tal analisi è stata condotta lungo due dimensioni: (a) quella delle funzioni di uso dell'applicaizone e di manutenzione del sistema, (b) quella della distribuzione fisica delle fonti dati/informazioni utilizzati dal sistema, già analizzata nel deliverable W3D1. Principali risultati di tale analisi sono: l'individuazione dei requisiti di un'architettura client/server dell'applicaizone che ne consenta il suo utilizza tramite il web (in sezione 3) e l'individuazione di alcune funzioni di manutenzione del sistema (in sezione 2
Discussing strategies for software and designing from an Agent-oriented point of view
No full textSoftware design is a complex problem-solving process which requires to evaluate several design options while pursuing the objective of adhering to general principles of good quality. For instance, designing open, robust and secure architectures for novel application areas, calls for evaluating and/or integrating different distributed system technologies, such as peer-to-peer and multi-agent systems.
This paper proposes a discussion on software architecting and designing strategies adopting an agent-oriented approach. Several works face this issue proposing methodologies based mainly on the use of architectural styles. Here we consider general strategies useful in the design process, such as divide-and-conquer, and principles for a
good design, such as cohesion and decoupling, and argue that, performing goal analysis, according to \emph{Tropos}, an agent-oriented methodology, provides methods for applying them. A case study in the domain of decision-support systems in agriculture is used to illustrat
Supporting Decision Making in Plant Disease Management
No full textThis paper describes a Decision Support System for plant disease management that has been designed for the use of the technicians of the Advisory Service of our region and of the experts who do research on disease management techniques, compliant with the Integrated Production and Organic Farming guidelines. AI methods and techniques that have been exploited during the development of the system are presented. In particular, we discuss the role of an Agent Oriented analysis of the application domain, for requirements elicitation and system design, and the exploitation of Machine Learning techniques in developing decision procedures at support of domain stakeholders when planning and executing actions for managing a plant disease. We illustrate results of their application with reference to a critical apple pest. The system architecture and the main user functions are also describe
Developing Tools for Agent-Oriented Visual Modeling
No full textMost of the Agent-Oriented software engineering methodologies that have been developed in the last years tend to propose a model-based approach to software development. To be put into practice, this approach demands flexible modeling tools as well as tools that provide transformation mechanisms to support the translation from one specification language to another, and finally to code, in a transparent and simple manner. The Model-Based Architecture initiative by OMG is proposing a challenging scenario for the research in this area and it is going to influence it providing standards. In this paper we describe a \ep which integrates an Agent-Oriented (AO) modeling tool with other tools, such as a model-checker for the verification of formal properties of the model and a library which implements graph transformation techniques which can be used to support model refinement as well as model transformations. Designing the tool we are taking into account reccomandations from the Model-Based Architecture initiative. A scenario is given with the aim of presenting practical motivations for supporting a light integration of different specification languages as well as the need for stronger synchronization mechanisms between different views on a model. We describe the \ep architecture, give details on the AO modeling tool and on the components that allows for the integration with other tool
- …
