Article Preview
Top1. Introduction
Due to the capability to effectively deal with changing demands and business environment, cyber-physical systems (CPS) emerge as a paradigm for manufacturers to respond to challenges in the real world and enhance system dependability as well as sustainability (Monostori et al., 2016). CPS are smart systems in which computational components in cyber space and physical components in physical space interact with each other to attain the objectives of sensing, monitoring and control (Graja et al., 2018). Reconfigurable Manufacturing Systems (RMS) provides a flexible architecture to accommodate frequent changes in CPS through dynamic configuration of resources (Koren et al., 1999; Son et al., 2001). A comprehensive literature review of the field of reconfigurable and intelligent machines is presented in (Molina et al., 2005). RMS can be regarded as a special class of CPS in the IoT infrastructure consisting of autonomous smart objects with the capabilities of sensing, processing, and networking (Kortuem et al., 2010). The complex interactions between workflows, machines and robots in the system presents several challenging research issues in the development of effective schemes to configure RMS. These research issues include modeling, formulation and optimization of RMS. An important issue in the design of CPS is to develop a mechanism for forming a team based on the available resources such as machines and robots in RMS. In the literature, although there are many studies on CPS, there lacks a framework to support dynamic process formation in CPS based on the concept of RMS. Motivated by the deficiencies in the existing studies, the goal of this study is to propose a solution methodology to achieve dynamic process formation for RMS by taking advantage of the state of art information technology.
Agent based manufacturing systems (Paolucci and Sacile, 2005) provide a decentralized, flexible and reconfigurable, manufacturing environment to accommodate changes and meet customers’ requirements dynamically. In the literature, a survey of the industrial application of agent technology in CPS can be found (Leitão at al., 2016). In agent-based manufacturing systems, manufacturing resources are represented by agents, which are autonomous, co-operative and intelligent entities able to collaborate with each other to process the tasks. The system of agents in an agent-based manufacturing system constitute a multi-agent systems (MAS) (Nilsson 1998; Ferber 1999). The objectives of production in agent-based manufacturing systems are achieved based on collaboration of a set of agents (Bussmann, Jennings and Wooldridge, 2004). In agent-based manufacturing systems, the concept of cooperative distributed problem solving (CDPS) is usually applied to determine whether agents can work together to solve problems that are beyond their individual capabilities based on coordination and cooperation (Durfee et al., 1989). The concept of cooperative distributed problem solving has been used in different problem domains in manufacturing sector. For example, (Hsieh, 2009; Hsieh & Lin, 2014; Hsieh & Lin, 2016; Ansari et al., 2018; Hsieh, 2018). In this article, we will apply the concept of cooperative distributed problem solving to solve the process formation problem for RMS. Given a set of heterogeneous agents, the problem is to form a team of agents to achieve the production goal. As there are many ways to achieve the production goal, this problem poses an optimization issue. To develop a systematic solution methodology, we will formulate this problem as a discrete optimization problem based on MAS architecture and focus on development of solution algorithms for this problem and assess the effectiveness of different solution algorithms.