Multi-objective hybrid algorithms for layout optimization in multi-robot cellular manufacturing systems

Highlights

• Hybrid algorithms to optimize layouts for multi-robot assembly systems are proposed.

• Layout area, operation time and manipulability are the three criteria used.

• DE, ABC, CSS and PSO are hybridized with genetic algorithm to have four hybrids.

• It is found that GA-PSO hybrid performs better over other hybrids.

• Algorithms are tested using Mitsubishi RV-6SQ robot configuration.

Abstract

Hybrid evolutionary algorithms to optimize layouts for multi-robot cellular manufacturing systems, which includes cooperative tasks among the robots is proposed in this paper. Layout area, operation time and manipulability of robot are the three design criteria useful to evaluate the robotic assembly systems are presented. Layout design candidates are represented using a sequence-pair scheme to prevent interferences between assembly system components, and the introduction of dummy components is proposed to represent layout areas where components are sparse. The main objective of this paper is to propose and evaluate hybrid algorithms by hybridizing them with genetic algorithm, which has been in use for decades. Differential evolution (DE), artificial bee colony (ABC), charged system search (CSS) and particle swarm optimization (PSO) are hybridized with genetic algorithm to have four hybrid (GA+DE, GA+ABC, GA+CSS and GA+PSO) algorithms. The performances of these algorithms are tested with genetic algorithm reported in the literature. The concept of non-dominated sorting genetic algorithm (NSGA-II) is borrowed to handle multiple objectives and to obtain Pareto solutions for the problems considered. These hybrid algorithms are evaluated using an example design problem of multi-robotic assembly system, and the effectiveness of these algorithms are presented in this paper. It is found that GA+PSO performs better over other hybrid algorithms considered. The application of the proposed algorithms are tested using Mitsubishi RV-6SQ robot configuration.

Introduction

In recent years, several researchers began investigating issues in multiple mobile robot systems. The field of cooperative autonomous mobile robotics is still new and this domain is not matured enough. Researchers working in this field are beginning to understand how to develop and control certain aspects of multi-robot systems [1]. The capabilities of Multi-robot systems are high when compared to single robot systems. But, there will be an increase in complexity of the control system to manage. A centralized control system which coordinates multiple robots may be very complex. When the robots do not share information, there is no need for any mechanism to update the global information; robots can be easily added to the system. The independent operation of individual robots can lead to increased flexibility and robustness since the addition of robots can be easily achieved without any need for changing software or hardware as more robots are added. However, optimal control cannot be achieved due to the lack of information [2].

Arai, et al. [1] identified the following seven principal topic areas of research in multi-robot systems:

• Biological Inspirations;

• Communication;

• Architectures, task allocation, and control;

• Localization, mapping, and exploration;

• Object transport and manipulation;

• Motion coordination; and

• Reconfigurable robots.

Object transport and manipulation are the key tasks occur in robotic assembly systems. These systems are also called as robotic cellular manufacturing system [18]. Enabling multiple robots to cooperatively carry, push, or manipulate common objects has been a long-standing, yet difficult, goal of multi-robot systems. Many research projects have dealt with this topic area; fewer of these projects have been demonstrated on physical robot systems. This research area has a number of practical applications that make it of particular interest for study [1].

Robots are widely applied in automated cellular manufacturing systems to transfer parts between stations, which are called as robotic cellular manufacturing systems (RCMS). RCMS are a type of manufacturing system in which one or more flexible robots carry out a large number of assembly operations that would be performed by human workers in conventional cellular manufacturing systems. When compared with conventional human cellular manufacturing systems, RCMS are seen to offer advantages, such as reduction of material flow distances and local inventory. However, although reduced operation costs due to automation of the manufacturing systems can be achieved by introducing RCMS, the design of assembly operations and robot teaching can become quite complex and time-consuming when launching new manufacturing systems. The layout design stage is one of the most important stages when seeking to build efficient RCMS. The layout design stage is an upstream stage of the manufacturing system design process, and decisions made then exert considerable influence on the detailed design of robot motion planning and the teaching process. Therefore, skillful decision-making during the layout design stage is essential to minimize design changes during the detailed design stage and teaching process, and to enhance the efficiency and reliability of the manufacturing systems when deployed [18].

The motivation of this paper is to propose multi-objective hybrid evolutionary algorithms to generate layouts for a multi-robot environment where assembly tasks are performed. In present day context robots are very much used in assembly systems for part transfer from storage to assembly stations and also between assembly stations.

This paper is organized as follows: Section 2 presents the literature review, the performance measures used to evaluate the layouts are discussed in Section 3, the implementation of sequence-pair representation scheme to generate layouts is explained in Section 4, the hybrid evolutionary algorithms proposed are presented in Section 5, Section 6 presents the conclusion and future research directions.