A modeling framework for manufacturing services in Service-oriented Holonic Manufacturing Systems

Highlights

• A framework for designing Manufacturing Services (MServices) is proposed.

• Service-oriented Holonic paradigm fosters flexibility in manufacturing control.

• MServices is the main communication primitive in distributed control systems.

• Automatic discovery of process capabilities is made possible with MServices.

• SoHMS formalism enables plug and produce and reusability principles.

Abstract

Holonic and Service-Oriented Architectures have been proposed as solutions for the conception of flexible and reactive systems. The combination of both architectures has been recognized as an attractive solution for the conception of more flexible and reactive systems. Services, originally conceived for web applications, need new models that better adapt to the needs of manufacturing systems namely, the specification and planning of processes. This paper proposes a methodology for designing customizable product-process specifications based on manufacturing-services that are suitable for product driven applications. A framework is proposed for modeling manufacturing-services designed for its application in Holonic Manufacturing Systems, preserving the fractal character found in products and processes. Three types of manufacturing processes are identified based on the relations among its composing operations and a model is proposed for each type. Application service-ontologies are created to describe operations and resource capabilities rather than complex semantic descriptions to facilitate process design and reduce the computational load of service discovery for planning. The framework leverages service reusability among process families in the same way as physical features are reused among product families. An illustrative example describes the design of a product's process specification based on the manufacturing service framework in a Holonic Manufacturing System, giving rise to Service-oriented Holonic Manufacturing Systems.

Introduction

An evolution in the goods market has been witnessed for the last few decades with a trend towards highly customized products and shorter product lifecycles. Such trend, expected to rise in the near future, is forcing companies on an exhaustive search for achieving responsiveness, flexibility, reduction of costs and increased productivity in their production systems in order to stay competitive in such new and constantly changing environment. The conception of “Next Generation Manufacturing Execution Systems” (NGMES), with the aforementioned attributes, has been challenging the community of Intelligent Manufacturing Systems, in a research guided by industry with the aspiration to incorporate “Agile Business” in manufacturing systems. Holonic Manufacturing Systems (HMS) and Service-Oriented Architectures (SOA) have been two of the most studied and referenced solutions, the former in manufacturing applications while the latter in informatics but making its way into industrial applications (such as in (Borangiu et al., 2012b)). Both of these solutions provide the necessary guidelines to create open, flexible and agile control environments for the NGMES.

The Holonic paradigm has been recognized, in industry and academics, as a provider of above mentioned attributes by means of a decentralized control architecture formed by a social organization of intelligent entities, called Holons, with specific behaviors and goals (Dehimi et al., 2015), defined by a reference architecture such as in (Chirn and McFarlane, 2000, Leitão and Restivo, 2006, Van Brussel et al., 1998). In another domain (computer science), the service-oriented paradigm defines the principles for conceiving decentralized control architectures that decompose computational processes into sub-processes, called services, to later distribute them among the different resources available (Rodríguez et al., 2016). Its focus is to leverage the creation of reusable and interoperable function blocks in order to reduce the amount of reprogramming efforts.

The combination of both paradigms appears to be a very attractive solution to the NGMES challenge as seen in works such as (Bellifemine et al., 2007, Jammes et al., 2005, Jammes and Smit, 2005) or more recently (Morariu et al., 2013a).This, thanks to the flexibility provided at a structural level of the control architecture and at a process level with the distribution and encapsulation of processes, was implemented using HMS and SOA respectively. Many recent works in literature have highlighted its benefits and proposed the use of services for the conception of control architectures and strategies for complex systems, such as the SOHOMA community (Borangiu et al., 2012b). However, until now, there has not been found works proposing a formal description of the informational elements forming and describing a service nor how this can form process specifications when integrated in the context of the manufacturing domain. Services, originally conceived for web applications, need new and more adequate models for their application in manufacturing systems. As services are the main element of negotiation in a service-oriented context, it is essential to define an appropriate model on which new architectures and interactions can be build up properly.

The design a Flexible Control System is a task that necessitates three successive steps: (i) the definition of an Information Model, (ii) the structural definition of the Control Architecture (defining the role of each entity in the architecture) and (iii) the definition of the Negotiation Mechanisms and/or Optimization Methods/Strategies enhancing the performance of the system. To obtain a truly agile control system, all these steps need to foster flexibility and reconfigurability. The objective of this work deals with the first of these issues: propose an information model based on the design principals of SoA adapted to the manufacturing context in order to respond to the interest of this domain such as interoperability, reusability and reconfigurability. Moreover, this adaptation is intended for its insertion in the classical Holonic Control Architectures such as PROSA (Van Brussel et al., 1998) that will provide process flexibility to the design of optimization strategies. The way these optimization tasks are defined and implemented is considered out of scope of this work. This work is motivated under the belief that the system's flexibility, being our main attribute of interest, is greatly determined by the way information is presented to the system, particularly at process specification. The objective of this paper is to propose a framework for designing Manufacturing Services (MServices) to form processes specifications that will allow the HMS's control architecture to explore the added process flexibility provided by the decomposition and encapsulation of process operations, now identified as MServices. These MServices represent an extension of the commonly known Processing Services, as those are related to some operations for processing or transforming the product in a manner determined by user-specific parameters (Sahin and Gumusay, 2008).

The paper is organized as follows. Section 2 reviews the existing process and service specifications from which terms and ideas are extracted. Section 3 presents the modeling framework describing what MServices are, and introduces the models of the MService perspectives. Section 3.2 revises the MService parameter model and how services are parametrized. Finally the methodology for process specification using the MService framework is described through an illustrative example in Section 4.