Some aspects of designing a dynamic cellular manufacturing system for deterministic and stochastic production requirements.

Jayakumar, V. ; Raju, R.

Introduction

Cell formation (CF) is a part of CM that is actually the implementation of group technology (GT) in manufacturing and production systems with the goal of classifying parts in a way that the physical or operational similarities of the parts are used in different aspects of design and production of parts. A cell formation problem (CFP) is the first stage in designing a CMS in order to form a set of manufacturing cells. A manufacturing cell consists of several functionally dissimilar machines which are placed in close proximity to one another and dedicated to the manufacture of a part family. Within each cell, a family of part types having similar manufacturing characteristics, physical features, production process information, product demand, and so forth are produced. Thus the tenet of CM is to break up a complex manufacturing facility into several groups of machines (cells), each being dedicated to the processing of a part family.

CM is a hybrid system linking the advantages of both job shops (flexibility in producing a wide variety of products) and flow lines (efficient flow and high production rate). As reported in the survey (Wemmerlov & Johnson, 1997), production planning and control procedures have been simplified with the use of CM. Obvious benefits gained from the conversion of the shop are: less travel distance for parts, less space required, and fewer machines needed. Since similar part types are grouped, this could lead to a reduction in setup time and allow a quicker response to changing conditions. In CM, machines are located in close proximity to one another and dedicated to a part family. This provides the efficient flow and high production rate similar to a flow line. The use of general-purpose machines and equipment in CM allows machines to be changed in order to handle new product designs and product demand with little efforts in terms of cost and time. So it provides great flexibility in producing a variety of products.

Design of Cellular Manufacturing Systems

Design of CMSs is a complex, multi-criteria and multi-step process. Given a set of part types, processing requirements, part type demand and available resources (machines, equipment, etc.,), the design of CMSs consists of the following three key steps: (i) part families are formed according to their processing requirements, (ii) machines are grouped into manufacturing cells, and (iii) part families are assigned to cells.

In the design of CMSs, design objective(s) must be specified. Minimizing intercell moves, distances, costs, and the number of exceptional parts (parts that need more than one cell for processing) are common design objectives. In addition to intercell material handling cost, other costs, such as machine cost, operating cost, etc., should be considered in the objective function in order to obtain more valid solutions. The design objective could be the minimization of the total of the sum of intercell material handling cost, equipment cost, and operating cost. Typical costs used in the design objective are: equipment cost, intercell material handling cost, inventory cost, machine relocation cost, operating cost, and setup cost. Costs in the design objectives may be conflicting; hence tradeoffs may need to be made during the design process. In addition to the design objectives, a number of strategic issues such as machine flexibility, cell layout, machine types, etc., need to be considered as a part of the CM design problem. Further, any cell configuration should satisfy operational goals (constraints) such as desired machine utilization, production volume, number of manufacturing cells, cell sizes, etc.

Comprehensive summaries and taxonomies of studies devoted to part-machine grouping problems were presented in Selim et al. (Selim, Askin and Vakharia, 1998), and Mansouri et al. (Mansouri, Moattar Husseini and Newman, 2000). In most of the articles reviewed by these authors and those published in recent years, cell formation problems have been considered under static conditions in which cells are formed for a single time period where product mix and demand are constant.

Dynamic Cellular Manufacturing Systems

Small make-to-order or subcontracting manufacturers produce a variety of parts for their customers in variable batch size; they must have a highly flexible but still competitive manufacturing production system. Because of the increasing variety of consumer goods and decreasing product life cycles, most manufacturing organizations encounter fluctuations in product mix and demand, known as dynamic or turbulent condition (Rheault, Drolet & Abdulnour, 1995). A turbulent environment is characterized by highly variable size of demand and production lots, highly variable processing times, highly variable setup times, partial or total stochastic demand, frequent changes of the product mix and/or unique order products, variable production sequences and a strong competition (Rheault et al, 1995). In such environment, there are a limited number of manufacturing system alternatives to improve the productivity and the efficiency of those manufacturers. At first, it seems that they are forced to adopt a traditional job shop layout, since group technology (GT) seems useless given the rapid ageing of classical cellular layout. But, a particular organization derived from cellular manufacturing does exist, it is the virtual cellular manufacturing system (VCMS). A virtual cell is a logical grouping of workstation that is not necessarily transposed into physical proximity. Rheault et al had proposed a framework for physically reconfigurable virtual cells by exploiting the moveableness of workstations. The CMS with a dynamic environment is referred to as dynamic cellular manufacturing system (DCMS). Most current cell formation methods have been developed for a single-period planning horizon. These methods assume that the problem data (e.g., product mix and demand) are constant for the entire planning horizon, in what is commonly known to as static environment (Jaydeep Balakrishnan & Chun Hung Cheng, 2007).

The product mix refers to the set of part types to be produced and the product demand is the quantity or volume of each part type to be produced. In a dynamic environment, the product mix and demand rate may vary under a multi-period planning horizon. In other words, this planning horizon can be divided into smaller time periods where each period and/or each part has different product mix and demand. A period can be a month, season, or year. As a consequence, the formed cells in the current period may not be optimal for the next period. This evaluation results from the reconfiguration of part families and machine grouping (Safaei, Saidi- Mehrabad, Tavakkoli-Moghaddam & Sassanic, 2008).

Reconfiguration consists of swapping the existing machines between cells (i.e., machine relocation), adding/removing machines to/from cells, replicating machines, and changing the process plan of the part-operation matrices. Chen (Chen, 1998) pointed out that cell reconfiguration, such as adding or removing machines, may be required in order to efficiently operate the system in a different manufacturing environment such as the electronic assembly industry. It is worth noting that reconfiguration may consist of increasing or decreasing the number of formed cells in each period.

Literature Review

The effectiveness of a cellular manufacturing system is sensitive to fluctuations in product demand and product mix (Chen, 1998; Song & Hitomi, 1996). The majority of existing cell formation models assume that both of these factors remain constant. In reality, the demand for a product varies over its life cycle, new products are introduced, and the production of older products is discontinued. Only a few models of the cell formation problem have attempted to address the changing nature of the production environment.

Chen (Chen, 1998) developed a mathematical programming model for system reconfiguration in a dynamic cellular manufacturing environment. He addressed the multi-period planning horizon with the introduction of new products and the discontinuation of old products. Song and Hitomi (Song & Hitomi, 1996) developed a methodology to design flexible manufacturing cells. The method integrates production planning and cellular layout via a long-run planning horizon. Wicks (Wicks, 1995) proposed a multi-period formation of the part family and machine cell (PF/MC) formation problem. The dynamic nature of production environment is addressed by considering a multi-period forecast of the product mix and resource availability during the formation of part families and machine cells.

Wilhelm et al (Wilhelm, Chiou & Chang, 1998) proposed a multi-period formation of the part family and machine cell (PF/MC) formation problem. Seifoddini (Seifoddini, 1998) addressed the uncertain nature of the product mix and proposed a probabilistic cell formation technique. Harhalakis et al (Harhalakis, Ioannou, Minis & Nagi, 1994) included random variations in product demand in their model. Unlike Seifoddini, who looked at multiple product mixes for a single period, Harhalakis et al. considered product demand changes over a series of periods in a planning horizon. However, neither of these models explicitly addressed the introduction of new parts to the production system. Vakharia and Kaku (Vakharia & Kaku, 1993) have developed a system redesign methodology that addresses long term demand changes. Schaller (Jeffrey Schaller, 2007) proposed an integer model that considers part reallocation or equipment reallocation between cells as alternatives for the redesign of a cellular manufacturing system to handle long-term demand changes. His model also considered the production cost within a cell based on the number of machines assigned to the cell. Balakrishnan (Balakrishnan & Cheng, 2007) conducted a comprehensive review of the work that addresses the modelling uncertainty and multi-period issues in CMS. They presented mathematical programming formulations as well as taxonomy of existing models.

There are two major drawbacks in current CM design methods. First is the lack of consideration of dynamic and uncertain production requirements. Second is the lack of accounting for the presence of routing flexibility in cellular systems, which often exists due to the availability of multi-function machines. The availability of routing flexibility presents alternative process plans to the system (Mungwatanna, 2000).

Dynamic and Stochastic Production Requirements

Most of the current CM design methods have been developed for a single-period planning horizon (static); they assume that problem data (e.g., product mix and demand) is constant for the entire planning horizon. Product mix refers to a set of part types to be produced, and product demand is the quantity of each part type to be produced. With shorter product life-cycles and time-to-market, it is likely that product mix and demand may change frequently (Chen, 1998). Wemmerlov and Johnson reported in their survey that the demand for finished products whose parts were manufactured in cells was not highly predictable. Therefore, a planning horizon can be divided into smaller periods where each period has different product mix and demand requirements. In such cases, we are faced with dynamic production requirements or a dynamic environment. Note that in a dynamic environment, product mix and/or demand in each period is different but is deterministic (i.e., known in advance).

In addition to dynamic production requirements, the product mix and/or demand in each period can be stochastic (uncertain), especially in future periods, due to customized products, shorter product life-cycles and unpredictable demand (Rheault et al, 1995). We refer to uncertainty of product mix and demand as stochastic production requirements or as a stochastic environment. In other words, stochastic pertains to not knowing exactly what changes in product mix and/or demand occur each period. The dynamic and stochastic nature of production requirements is independent.

In employing a single-period planning horizon with known demand, current CM design methods assume a static, deterministic environment. This approach, however, could decrease the validity of any CM solutions so obtained. With frequent changes in the product mix and/or demand, manufacturing cells must be modified from time to time. The optimal cell configuration generated from earlier data may not be valid after such changes occur in the system (Mungwatanna, 2000). Intercell material handling costs could increase after such changes. In order to successfully utilize CM in such environments, the original system must evolve over time to match the changing conditions. This evolution results from reformulation of part families, manufacturing cells and reconfiguration of the cellular manufacturing system as required. Reconfiguration consists of swapping existing machines between cells, adding new machines to cells, removing existing machines from cells, and/or replacing existing machines in cells (Safaei et al, 2008).

Unfortunately, few works in the design of CMSs have addressed dynamic and stochastic production requirements. However, they have gained interest from researchers in recent years (Chen, 1998; Harhalakis et al, 1994; Seifoddini, 1998; Sethi & Sethi, 1990; Song & Hitomi, 1996; Vakharia & Kaku, 1993). Certain CM design strategies have been suggested to deal with dynamic production requirements. A robust design strategy is to design a cellular manufacturing system that is good for the entire planning horizon even though it may not be optimal in any period. An adaptable design strategy is to design a CM that responds to changing product mix and/or demand in future periods by rearranging the current manufacturing system. By rearranging the system, it is hoped that the reduction in material handling costs will offset the rearrangement costs (Wicks, 1995).

One of the efforts partially addressing the problem of dynamic and stochastic production requirements is called "agile manufacturing". Agile manufacturing refers to manufacturing utilizing resources and people that can be changed, or reconfigured, quickly and easily to cope with variability and uncertainty. The concept of agile manufacturing is motivated by the unpredictability in demand and how it can be handled (Mungwatanna, 2000).

Conclusion

Summarizing the above we see that, despite the growing importance of CMSs, available methods for designing CMSs provide generally a piecemeal approach when formulating CMS models. They do not consider in an integrated manner several important factors. These factors are the dynamic and stochastic nature of production requirements, the availability of routing flexibility, lot splitting, operation sequence, duplicate machines, machine capacity and machine procurement. By considering these factors, CM design solutions can be improved. A mathematic model may be developed to capture those requirements. The main objective of this paper is to address the need of a design method that can integrate accurately and realistically the existence of dynamic and stochastic production requirements and production planning aspects during the CMS design. Therefore more research work may be attempted to develop a design methodology for cellular manufacturing systems in dynamic and stochastic production environments which employs system-dependent reconfigurations and routing flexibility. Also due to the fact that CF is a NP- hard problem, then solving the model using classical optimization methods needs a long computational time; therefore the use of meta-heuristics such as Tabu Search, Genetic Algorithm, and Simulation Annealing Algorithm may be developed and investigated especially to solve problems of larger scale.

References

[1] Wemmerlov, U., and Johnson, D., 1997, "Cellular Manufacturing at 46 User Plants: Implementation Experiences and Performance Improvements", International Journal of Production Research, 35(1), pp. 29-49.

[2] Selim, H.M., Askin, R.G., and Vakharia, A.J., 1998, "Cell Formation in Group Technology: Review Evaluation and Direction for Future Research", Computers and Industrial Engineering, 34, pp. 2-30.

[3] Mansouri, S.A., Moattar Husseini, S.M., and Newman, S.T., 2000, "A Review of The Modern Approach to Multi-Criteria Cell Design", International Journal of Production Research, 38, pp. 1201-1218.

[4] Rheault, M., Drolet, J., and Abdulnour, G., 1995, "Physically Reconfigurable Virtual Cells: A Dynamic Model for a Highly Dynamic Environment", Computers and Industrial Engineering, 29(4), pp. 221-225.

[5] Jaydeep Balakrishnan., and Chun Hung Cheng., 2007, "Multi-period Planning and Uncertainty Issues in Cellular Manufacturing: A Review and Future Directions", European Journal of Operational Research, 177, pp. 281-309

[6] Safaei, N., Saidi-Mehrabad, M., Tavakkoli-Moghaddam, R., and Sassanic, F., 2008, "A Fuzzy Programming Approach for a Cell Formation Problem with Dynamic and Uncertain Conditions", Fuzzy Sets and Systems, 159, pp. 215-236.

[7] Chen, M.. 1998, "A Mathematical Programming Model for System Reconfiguration in a Dynamic Cellular Manufacturing Environment", Annals of Operations Research, 77(1), pp. 109-128.

[8] Song, S., and Hitomi, K., 1996, "Integrating the Production Planning and Cellular Layout for Flexible Cellular Manufacturing", Production Planning and Control, 7(6), pp. 585-593.

[9] Wicks, E., 1995, "Designing Cellular Manufacturing Systems with Time Varying Product Mix and Resource Availability", PhD thesis, Virginia Polytechnic Institute and State University, Blacksburg, VA.

[10] Wilhelm, W., Chiou, C., and Chang, D., 1998, "Integrating Design and Planning Considerations in Cellular Manufacturing", Annals of Operations Research, 77(1), pp. 97-107.

[11] Seifoddini, H., 1990, "A Probabilistic Model for Machine Cell Formation", Journal of Manufacturing Systems, 9(1), pp. 69-75.

[12] Harhalakis, G., Ioannou, I., Minis, I., and Nagi, R., 1994, "Manufacturing Cell Formation under Random Product Demand", International Journal of Production Research, 32(1), pp. 47-64.

[13] Vakharia, A., and Kaku, B., 1993, "Redesigning a Cellular Manufacturing System to Handle Long-Term Demand Changes: A Methodology and Investigation", Decision Sciences, 24(5), pp. 84-97.

[14] Jeffrey Schaller., 2007, "Designing and Redesigning Cellular Manufacturing Systems to Handle Demand Changes", Computers & Industrial Engineering, 53, pp. 478-490,

[15] Mungwatanna, A., 2000, "Design of Cellular Manufacturing Systems for Dynamic and Uncertain Production Requirement with Presence of Routing Flexibility", Ph.D. Thesis, Blacksburg State University Virginia.

[16] Sethi, A., and Sethi, S., 1990, "Flexibility in Manufacturing: A Survey", International Journal of Flexible Manufacturing Systems, 2, pp. 289-328.

(1) V. Jayakumar and (2) R. Raju

(1) Assistant Professor, Department of Mechanical Engineering, Velammal Engineering College, Chennai-600 066, Tamil Nadu, INDIA Corresponding Author E-mail: jkmails2k2@yahoo.com

(2) Assistant Professor, Department of Industrial Engineering, Anna University Chennai, Chennai-600 025, Tamil Nadu, INDIA

Authors Identification Note

Sri V. Jayakumar is working as Assistant Professor at the Department of Mechanical Engineering, Velammal Engineering College, Chennai, Tamil Nadu. He is currently pursuing hid PhD in College of Engineering, Guindy Campus, Anna University- Chennai. He has more than 10 years of Teaching Experience. He has published 13 Research Papers in International & National Conferences, and 5 Textbooks on Mechanical Engineering for Engineering Graduate Students of Anna University.

Dr. R. Raju is working as Assistant Professor at the Department of Industrial Engineering, College of Engineering, Guindy Campus, Anna University--Chennai, Chennai, Tamil Nadu. He has more than 25 years of Industrial Experience and five years of Teaching Experience. He has published four articles in International Journals and about fifteen in National journals. His area of interest is Industrial Engineering in general and Quality Engineering and Management in particular.