Bio-inspired multi-agent systems for reconfigurable manufacturing systems☆
Introduction
The current global economy imposes new challenges for manufacturing companies, with cost, quality and responsiveness being the three critical foundations on which every manufacturing company stands to remain competitive (ElMaraghy, 2006). Under these circumstances, manufacturing systems are required to be more flexible, robust and reconfigurable, supporting the agile response to the changing conditions through their dynamic re-configuration on the fly (i.e., without stopping, re-programming or re-starting the processes or the other system components). Since the systems are more complex, distributed and reconfigurable, the probability of the system malfunction also increases (Trentesaux, 2009).
Since the traditional approaches, based on centralized, rigid structures, do not have enough flexibility to cope with modularity, flexibility, robustness and re-configuration, several paradigms have been introduced over the last few years: Multi-Agent Systems (MAS) (Wooldridge, 2002), Holonic Manufacturing Systems (HMS) (Deen, 2003, Leitão and Restivo, 2006) and Bionic Manufacturing Systems (BMS) (Okino, 1993). In spite of their natural differences, these paradigms propose distributed, autonomous and adaptive manufacturing systems, which can respond promptly and correctly to external changes. These paradigms differ from the conventional approaches due to their inherent ability to adapt to changes without external interventions. In addition, the HMS and BMS paradigms indicate that hierarchy is needed to guarantee the inter-entity conflict resolution and to maintain overall system coherence and objectivity in the face of the individual, autonomous attitude of the entities (Sousa et al., 1999).
The work on MAS, HMS and BMS provides a good framework to rise to the challenge of developing a new class of adaptive and reconfigurable manufacturing systems that will support robustness and re-configurability quite naturally (see the surveys of Leitão, 2009a, Pechoucek and Marik, 2008). However, the current application of such paradigms usually does not consider self-adaptation and self-organization, which results in the systems becoming increasingly reconfigurable, adaptive, organized and efficient.
In biology and nature, complex systems behave simply because of the cooperation of individuals, who are very simple, with very limited cognitive skills (e.g., the colonies of ants and bees). Biological insights have been the source of inspiration for the development of several techniques and methods to solve complex engineering problems, such as logistics and traffic optimization, telecommunications networks, economic markets and production systems (Leitão, 2009b). Multi-Agent Systems have already inherited certain ideas derived from biology and nature, but they can be enhanced with other biological insights, notably the swarm intelligence and self-organization, to obtain more responsive, adaptive systems that address the current requirements imposed on manufacturing systems. In particular, bio-inspired techniques can contribute to obtaining manufacturing systems with the desired characteristics of flexibility, robustness, re-configuration and responsiveness.
The motivation of this paper is to understand how bio-inspired techniques can be used to solve complex engineering problems. Thus, some biological phenomena are studied, some of the existing bio-inspired applications are analyzed, and then the real benefits of bio-inspired MAS for solving the current manufacturing control problems are examined. A real implementation of a bio-inspired solution for routing pallets in a flexible manufacturing system is described to illustrate its suitability.
The remainder of this paper is organized as follows. Section 2 provides an overview of biological phenomena, and Section 3 introduces some bio-inspired techniques and methods used to solve complex problems, especially manufacturing problems. Section 4 discusses the suitability of bio-inspired multi-agent systems solutions for the manufacturing field, and Section 5 describes a bio-inspired solution based on potential fields for an industrial automation case study. Finally, Section 6 presents our conclusions and our prospects for future research.
Section snippets
Basic concepts found in biology
Nature offers plenty of powerful mechanisms, refined by millions of years of evolution, to handle emergent and evolvable environments (Leitão, 2009b). This section tries to show how complex things behave simply in nature and biology, introducing the concepts of swarm intelligence (Section 2.1) and evolution and self-organization (Section 2.2).
Survey of bio-inspired applications for solving complex problems
Several researchers used biological behavior (e.g., colonies of insects) to solve complex mathematical engineering problems. In this section, bio-inspired techniques and methods in engineering are briefly reviewed, with special attention to their applicability in manufacturing.
Applicability of bio-inspired systems in manufacturing
The analysis in the previous section shows the tremendous potential of using of bio-inspired systems to solve complex engineering problems. This section discusses the applicability and benefits of combining bio-inspired techniques with multi-agent systems in the manufacturing domain in order to address the current challenges.
The MAS paradigm has already inherited biological insights (Barbosa and Leitão, 2010):
- •
Distributed nature—multi-agent systems are based on a set of distributed, autonomous
An automation case study: a bio-inspired approach
An experimental case study was used to demonstrate the applicability of bio-inspired multi-agent systems in manufacturing. For this purpose, the full-size flexible manufacturing system (FMS), located at the AIP-PRIMECA production center (Université de Valenciennes et Hainaut-Cambrésis, France), was used. The FMS is composed of seven work stations (i.e., assembly robots, quality control units and load/unload units), interconnected by a flexible conveyor system that uses shuttles to move pallets
Conclusions
This paper analyzed some mechanisms found in biology and nature, especially swarm intelligence and self-organization, and tried to understand their potential benefits to solve complex engineering problems. Special attention was devoted to the existing bio-inspired applications, particularly in manufacturing. This paper also discussed the application of bio-inspired techniques to enhance multi-agent systems in the different manufacturing areas and considered how to achieve a greater adoption in
Acknowledgments
The experimental work was performed by a project team, including Nadine Zbib, Cyrille Pach, Yves Sallez, Thierry Berger, with the help and support of the AIP-PRIMECA team. The authors wish to thank all of them for their work.
References (176)
- et al.
A particle swarm optimization for the vehicle routing problem with simultaneous pickup and delivery
Comput. Oper. Res.
(2009) - et al.
Ant colony optimization techniques for the vehicle routing problem
Adv. Eng. Inf.
(2004) - et al.
Max–min adaptive ant colony optimization approach to multi-UAVs coordinated trajectory replanning in dynamic and uncertain environments
J. Bionic Eng.
(2009) - et al.
Scheduling continuous casting of aluminum using a multiple objective ant colony optimization metaheuristic
Eur. J. Oper. Res.
(2002) Agent-based distributed manufacturing control: a state-of-the-art survey
Int. J. Eng. Appl. Artif. Intell.
(2009)- et al.
ADACOR: a holonic architecture for agile and adaptive manufacturing control
Comput. Ind.
(2006) - et al.
Particle swarm optimization for pap-smear diagnosis
Expert Syst. Appl.
(2008) - et al.
A market approach to holonic manufacturing
Ann. CIRP
(1996) - Abd-El-Barr, M., Sait, S., Sarif, B., 2003. Ant colony algorithm for evolutionary design of arithmetic circuits. In:...
- Aggoune, R., Mahdi, A., Portmann, M., 2001. Genetic algorithms for the flow shop scheduling problem with availability...
Centralized versus decentralized control in manufacturing: lessons from social insects
Complexity Complex Syst. Ind.
An ant colony optimization algorithm for shop scheduling problems
J. Math. Modelling Algorithms
Swarm Intelligence: from Natural to Artificial Systems
Self-organization in social insects
Trends Ecol. Evol.
Multiagent Systems for Manufacturing Control
Self-Organisation in Biological Systems, Princeton Studies in Complexity
Artificial bee colony algorithm for transient performance augmentation of grid connected distributed generation
LNCS
Ant colony optimisation for machine layout problems
Comput. Optim. Appl.
The self-organizing exploratory pattern of the Argentine ant
J. Insect Behav.
AntNet: distributed stigmergetic control for communications networks
J. Artif. Intell. Res.
Bumble bees alert to food with pheromone from tergal gland
J. Comp. Physiol. A: Neuroethol. Sens. Neural Behav. Physiol.
Flexible and reconfigurable manufacturing systems paradigms
International Journal of Flexible Manufacturing Systems
Cited by (140)
Distributed resource scheduling in edge computing: Problems, solutions, and opportunities
2022, Computer NetworksReprint of: Gentelligent processes in biologically inspired manufacturing
2021, CIRP Journal of Manufacturing Science and TechnologyReprint of: Application cases of biological transformation in manufacturing technology
2021, CIRP Journal of Manufacturing Science and TechnologyCitation Excerpt :Ueda et al. introduced the concept of Biological Manufacturing Systems in 1997 and studied self- organizing manufacturing systems as a solution to dynamic changes in external and internal environments [6]. More recent research on self-organizing manufacturing systems can be found in Ref. [7]. Biologically inspired design has been investigated in Ref. [8].
Adaptable automation with modular deep reinforcement learning and policy transfer
2021, Engineering Applications of Artificial Intelligence
- ☆
Note: This paper is an extended version of a paper presented at the 10th IFAC Workshop on Intelligent Manufacturing Systems (IMS’10), selected by the organization to be submitted to the EAAI Journal.