Catálogo de publicaciones - libros

In this paper, we present a generative design method for supporting creative conceptual design in a multi-agent system. The hierarchical multi-agent system architecture offers a promising framework and an evolutionary computational approach provides a basis for supporting generative design in distributed environments. The evolutionary approach relies upon a tree structure rather than a known binary string in general GA (Genetic Algorithm). The system can generate simple 2D sketch shapes, which are generated by using general mathematical expressions, and selected shapes are dealt with to form 3D components. These components are classified and saved in a SQL sever based database (component base). The complex design is implemented via combination of different components that come from the component base. The layout of components is generated by a genetic algorithm. The generative process is illustrated through an architectural design example.

In this paper, we start from a case study of an architecture company and move on to a framework of agents to identify and inform conceptual design. The processes and problems exhibited by the company involved are quite common and can be found in other similar companies. Our intent is to instrument the current process using agent technology, in an effort to improve global awareness during the conceptual design phase, allowing the designers to design so as to facilitate later steps and optimize the process as a whole.

New design optimization techniques and advanced computer technologies have laid the foundation for the emerging fields of distributed multidisciplinary design optimization (MDO). The challenge now faced by designers is to effectively use this vast amount of knowledge. There are many software tools available in each discipline. However, the key problem is how to integrate these tools and databases, which conform to different interfaces and requirements with little consideration to integration and reusability, in a flexible and robust way. This paper proposes a novel ontology-based semantic integration framework for cooperative distributed MDO environments. The semantic integration aspect will provide the foundation for service-oriented architecture for distributed MDO environments. The cooperation aspect will focus on seamless integration among autonomous MDO systems in dynamic open environments, using multi-agent paradigm.

This paper presents a model of formal dialogue between intelligent and autonomous agents and its application to team formation in cooperative design. Using formal dialogue, agents exchange information about their specialty domain, ability, opportunity and willingness. Agents understand each other to form a potential cooperative team, in which joint beliefs and intentions are created. The cooperative tasks can be carried out within the collaborative team, and avoid the blindness in the task decomposition and allocation as usually appeared in cooperative design systems.

Both CORBA and mobile agent technologies have caught a wide attention in research and development fields. Integrating the two technologies to make CORBA objects movable is the main idea of this paper. A CORBA based mobile agent server architecture, called MA_CORBA, is proposed and its prototype is presented.

Focusing on exceptions that occur frequently in Computer Supported Cooperative Design (CSCD), the definition and classification of the exceptions are presented. According to expected and unexpected characteristics of exceptions, three methods are proposed: (1) expected exceptions are automatically dealt with by using Agent Message Event Rule (AMER) during the execution of collaboration processes; (2) document related exceptions were dealt with by adopting Document Tracking Log; (3) the cause of unexpected collaboration exceptions (UCE) is analyzed by using the algorithm of similarity-matching based on knowledge for mining exception cases in order to get the solution of similar exceptions. A prototype system, CoopDesigner, based on the proposed methods, is presented at the end of the paper.

An increasing number of sites including scientific electronic journals, digital libraries and collaboratories have been dedicated to scientific knowledge sharing and exchange in the recent years. Despite the apparent progress, in which traditional journals are migrating to Web and new forms of communication emerge, a closer look at these sites indicates that the scientific publishing model remains essentially unchanged. In this context was born the Configurable Electronic Journal (CEJ) project, a solution for generating configurable and extensible electronic journals. One of the main goals of this work is to apply open source technologies, including W3C Semantic Web technologies, in the design of an innovative environment for the definition of publication processes. CEJ environment is flexible to accommodate different styles of publications, suitable to be used in different processes and disciplines.

This paper first provides a literature review of product development and knowledge management including product development process, design history and domain knowledge. It then presents a method for knowledge-based multi-view process modeling including process implementation, process monitoring, and knowledge management. The integrated framework and hierarchical model of the design history is built. The relationship among process history, design intent and domain knowledge is analyzed, and the method for representation and acquisition of process history, design intent and domain knowledge is presented. The multi-agent based architecture of the knowledge-based product development process management (PDPM) system and its functional modules are described, and the architecture of integrated knowledge management (IKM) system based on PDPM is set up and developed with a B/C/S (Browser/Client/Server) structure. The system is used successfully during the lifecycle of a new type of railway rolling stock development.

According to the requirements of cooperative product development in networked manufacturing environments, a knowledge-based cooperative design technology is proposed. Based on the Chengdu-Deyang-Mianyang Networked Manufacturing and ASP Platform, a knowledge-based networked manufacturing cooperative design platform is developed by exploiting the product function modules, parts management, supplier management, customer management, and resources allocation optimization. In the cooperative design platform, the integration of manufacturing resources of supply chain and the method of client customization and enterprise cooperation are realized in the course of product development. The platform has been put into application in a number of enterprises in the Chengdu-Deyang-Mianyang region of China.

Online configuration for products in the distributed and dynamic computing environment motivates the demand for semantic based cooperation, which takes place in the supply chain under B2B situation. Traditional standalone knowledge model of configuration systems does not meet the new requirements. We propose a multi-ontology based solution. The core idea of this paper is to take the process knowledge of distributed systems into consideration, which supports the integration among distributed configuration systems. Furthermore, the process model provides possibility to optimize solutions for configuration problems. OWL is used as modeling language in order to utilize potential benefits of current Semantic Web technology.