EMRS | alumni

Surface mount Printed Wiring Board (PWB) assembly requires a variety of assembly processes and automated machinery to manufacture a range of PWB types. Throughput in PWB assembly can be improved by effectively exploiting the utilization and integration of the assembly equipment and processes. Conventionally, production planning in a multiple assembly line PWB assembly environment was implemented either by human planners or by stand alone decision support system. However, with the increasing complexity and density of electronic packages and the short life cycle of electronic products, real time process planning is required to dynamically deal with the problems faced in the PWB assembly environment. Also, current decision support systems, which include manufacturing systems and stand alone knowledge based systems, have difficulty in effectively communicating with each other to arrive at a common solution(s). The global objective of this research was to exploit the integration and coordination that is possible in a surface mount PWB assembly facility through the use of DAI technology. The developed DAI prototype system helps to provide real-time responses to manufacturing related problems. This DAI based system allows for the design and implementation of an efficient mechanism to deal with dynamic production and process planning for PWB assembly.

DAI technology in the manufacturing domain, especially in the field of surface mount technology, has not been widely explored. This research, which focuses on a DAI framework to solve production planning problems by using multiple expert systems, takes the first step in tlis direction. This system is unique as it not only utilizes the data transmission techniques existing in the traditional distributed manufacturing systems approach but also incorporates multiple knowledge based systems that work together to solve problems in the domain of interest.

The prototype DAI system utilizes a group of problem solvers (intelligent agents) to form a problem solving team which implements dynamic production planning and production control in the problem domain. This prototype system was developed and implemented on a Local Area Network (LAN) where Personal Computers (PCs) are connected using bus topology. Each PC is considered as an intelligent agent in the DAI system. The DAI system was developed using Object Oriented Programming (OOP) concept which consider each intelligent agent to be an object in the software. The overall implementation of the system was then achieved by the manipulation of the individual objects/agents. This design concept allows further enhancement of the system by adding new objects (agents) to the system. An OOP development tool, Borland C ++, was used for encoding all the reasoning, calculation, and communication processes required by the DAI system. This development tool incorporates the Microsoft Windows graphical user interface, thus providing a user friendly environment. The prototype system consists of six intelligent agents which are used to implement the dynamic production planning task. The planning task is divided into sub-tasks which can be solved by the individual intelligent agents with respect to their knowledge domains. The communication and cooperation between the intelligent agents is achieved by exchanging the messages and information within the blackboard communication framework. To resolve the conflict between various knowledge domains of the intelligent agents, a fuzzy coordination technique was used to negotiate the various solutions generated by the intelligent agents. This distributed processing capability allows intelligent agents to solve problems independently and concurrently, simplifying complex problems and providing instant decision support.


	sitemap \| contact us