Genetic vs. Hybrid Algorithm in Process of Cell Formation

Genetic vs. Hybrid Algorithm in Process of Cell Formation

R. Sudhakra Pandian (Kalasalingam University, India), Pavol Semanco (Technical University of Kosice, Slovakia) and Peter Knuth (Technical University of Kosice, Slovakia)
DOI: 10.4018/978-1-61350-047-7.ch005
OnDemand PDF Download:
List Price: $37.50


The cell formation problem has met with a significant amount of attention in recent years by demonstrating great potential for productivity improvements in production environment. Therefore, the researchers have been developing various methods based on similarity coefficient (SC), graph theory approaches, neural networks (NN), and others with aim to automate the whole cell formation process. This chapter focuses on presentation of hybrid algorithm (HA) and genetic algorithm that are helpful in production flow analysis to solve the cell formation problem. The evaluation of hybrid and genetic algorithms are carried out against the K-means algorithm and C-linkage algorithm that are well known from the literature. The comparison uses performance measure and the total number of exceptional elements (EEs) in the block-diagonal structure of machine-part incidence matrix using operational time as an input. The final performance results are presented in the form of graphs.
Chapter Preview


In recent years, the cell formation problem has received a significant amount of attention by demonstrating a great potential for productivity improvements of cellular manufacturing system (CMS), which groups machines with dissimilar function and workstation types, dedicated to family of similar components. The main problem in designing of cellular manufacturing system is the formation of part families and corresponding groups of machines. One of the methods for classification of cell formation is production flow analysis (PFA). The concept of PFA, proposed by Burbidge (1977)is probably the most well-known and most widely accepted. This method requires reliable and well-documented route sheets and is also time-consuming. The researchers have initiated development of various methods like similarity coefficient method, graph theoretic approaches, array based methods, etc. in this field with aim to automate the cell formation process.

The modelling of CMS through mathematical programming was started to incorporate more real life constraints on the problem. Later researchers began developing heuristics and meta-heuristics methods to explore the best optimal solutions for the Cell Formation (CF) problems. Since soft computing techniques nowadays expand their applications to various fields like telecommunications, networking, design and manufacturing, current research in CMS is being carried out using soft computing techniques.

Very few studies focus on CF considering production factors such as operational time, operational sequence, batch size etc. In this chapter some of the real-time production factors are considered. For this purpose, the zero-one binary machine part incidence matrix (MPIM) of CF problem is converted into real valued operational time data. The use of soft computing technique is found more suitable for such type of problems, because it is capable of producing reliable results.

One of the chapter objectives is introduction of heuristic and meta-heuristic approaches based on similarity coefficient for solving cell formation problem. Simultaneously some important production factors that cannot be ignored during cell formation are presented. Another objective is to propose suitable methodologies for cell formation considering real time production factors using hybrid and genetic algorithms. The evaluation of the particular algorithms will be carried out by use of modified grouping efficiency (MGE) as a measure of performance.

Complete Chapter List

Search this Book: