A performance evaluation of industrial agents

Assembly systems today face significant pressure to provide highly adaptable and quickly deployable solutions in order to deal with unpredictable changes according to market trends. However, control of assembly processes is dominated by the use of Programmable Logical Controllers (PLC) which do not provide the necessary mechanisms to easily deal with these challenges. The concept of agent-based control has been introduced as a solution to deal these challenges and support new production paradigms based on the plug and produce concept. However, this solution has not yet been proven to be a real alternative to the traditional PLC approach in terms of performance. This work is investigating the use an approach that is able to benefit from the relative advantages of both PLC and agents solutions. A new hybrid architecture is presented which combines the functionalities of a PLC with those of industrial agents. The focus is on assessing the performance of this approach. The proposed approach is applied to an assembly station using only the core elements of the model. A number of control strategies to coordinate presses steps in the assembly sequence are compared to a pure PLC control approach. First results indicate that it is possible to achieve comparable response times with the hybrid approach. The outcomes of this work will provide the basis for further refinement of the proposed hybrid agent architecture and is intended to lead to the definition of a much needed more comprehensive assessment framework for industrial agent systems in the production systems context.