Int. J. Production Economics 85 (2003) 141–153
Flexibility configurations for the supply chain management A. Claudio Garavelli* DIMEG, Politecnico di Bari, Viale Japigia 182, 70126 Bari, Italy
Abstract Facing a dynamic and complex environment, networked companies often require the coordination of many plants, which produce and deliver goods to customers located in different places, and suppliers, which provide each plant with the required components. A way to optimize the product flows in supply chains (SCs) is to adopt the concept of limited flexibility, that is a particular configuration of product assignments to plants and components to suppliers which can yield many benefits without dramatically increasing the flexibility costs. In this paper, a simulation model is proposed to evaluate the performance of different configurations of a SC. In particular, based on a work-in-process and time performance analysis, the different configurations are analyzed in order to support the selection of suitable flexibility degrees of the operations network. r 2003 Elsevier Science B.V. All rights reserved. Keywords: Supply chain; Operations network; Limited flexibility; Simulation; Performance
1. Introduction Nowadays, many businesses move human resources, materials and information from some place to some other place throughout the world. From this point of view, these ‘‘network-based’’ businesses share a common structure (Coyne and Dye, 1998). Operations flexibility can be considered a crucial weapon to increase competitiveness in such a complex and turbulent marketplace (Upton, 1994). Flexibility becomes particularly relevant when the whole supply chain (SC) is considered, consisting of a network of supply, production, and delivering firms (Christopher, 1992). In this case, many sources of uncertainty have to be handled, such as market demand, supplier lead time, product quality, and informa*Tel.: +39-080-5962-719; fax: +39-080-5962-788. E-mail address:
[email protected] (A.C. Garavelli).
tion delay (Giannoccaro et al., 2003). Flexibility allows to switch production among different plants and suppliers, so that management can cope with internal and external variability (Chen et al., 1994). In a global scenario, not only manufacturing, but also logistics can be an important source of competitive advantage, since material flows strongly affect business performance. Different logistics channels of the SC can be activated in order to face emergencies such as demand peaks. Management is continuously involved in decision making, often without the adequate informative support, thus causing frequent shifts of workloads within the system that increase production delay and transport cost (Egbelu, 1991). The production order assignments to the plants and the organization of transports are then critical decisional factors that can decrease the performance of a wide range of products (Albino et al., 2002).
0925-5273/03/$ - see front matter r 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0925-5273(03)00106-3
142
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
A strategy of coordination of the different production plants is required to address the challenge of several markets (Bhatnagar et al., 1993; Kogut and Kulatilaka, 1994). To coordinate the production network it is necessary to analyze the flexibility of the system components and their relationships, in order to evaluate their impact on the whole system. Despite flexibility and SC management have been among the leading concerns of operations managers for several years, there are not many specific studies on the SC flexibility in the literature (Duclos et al., 2001). This paper intends to offer a contribution in this direction. In particular, a model is proposed for the evaluation of the effects of different degrees of flexibility on the performance of a multi-product SC, subject to stochastic market demands and made up of several customers, assembly facilities, and supplier sites. In this way, the design of the SC configuration, which is one of the critical problems in SC management, can be investigated (Giannoccaro and Pontrandolfo, 2001). In the literature, there are some studies addressed to evaluate the effect of flexibility configurations on SC planning performance, such as lost sales (e.g., Albino et al., 2002). In this paper, similarly to other simulation studies focused on the flexibility of a single-stage production system (e.g., Garavelli, 2001), SC lead time and work-inprocess are investigated. The paper is organized as follows. In Section 2, the type of SC flexibility investigated in the paper is pointed out. In Section 3, different SC flexibility configurations are described. In Section 4, a simulation model is defined for the performance evaluation of the SC flexibility configurations. In Section 5, the performance analysis of a SC characterized by three suppliers, assemblers, markets, and product types is carried out. Final considerations are reported in Section 6.
2. Flexibility in a SC As commonly shared by the literature on manufacturing systems, flexibility is a complex and multidimensional concept, difficult to sum-
marize (Upton, 1994; Gupta and Buzacott, 1996). According to a broad definition, flexibility reflects the ability of a system to properly and rapidly respond to changes, coming from inside as well as outside the system. Referring to the several papers which have proposed useful taxonomies, different aspects of flexibility can be outlined, such as functional aspects, i.e. flexibility in operations, marketing, logistics, etc. (e.g., Kim, 1991; Lynch and Cross, 1991), hierarchical aspects, i.e. flexibility at shop, plant or company level (e.g., Gupta, 1993; Koste and Malhotra, 1999; Slack, 1987), measurement aspects, focused on global flexibility measures vs. context specific ones (e.g., Gupta and Somers, 1992; Sarker et al., 1994; Chung and Chen, 1990; de Groote, 1994), strategic aspects, centered on the strategic relevance of flexibility (e.g., Chambers, 1992; Gerwin, 1993; Nakane and Hall, 1991), time horizon aspects, e.g. long-term vs. short-term flexibility (Zelenovich, 1982). From an operational perspective, however, the most interesting aspect of flexibility is probably the one concerning the object of change, i.e. flexibility of product, mix, volume, etc. (e.g., Gerwin, 1993; Browne et al., 1984; D’Souza and Williams, 2000; Hyun and Ahn, 1992; Sethi and Sethi, 1990; Vokurka and O’Leary-Kelly, 2000). Although the papers available on the specific subject of SC flexibility are not very numerous, it is possible to find some definitions of the various types of SC flexibility, usually associated to correspondent types of flexibility of manufacturing systems and referred to the object of change. For instance, Vickery et al. (1999) propose the following dimensions of SC flexibility: product, volume, launch, access, target market, while Viswanadham and Srinivasa Raghavan (1997) consider: volume, mix, routing, delivery time, new product flexibility. The analysis of the SC flexibility involves the consideration of the flexibility of the SC components and their relationships, in order to evaluate their impact on the whole system. In this paper, a two-stage SC is considered. For the sake of simplicity, the actors of the SC are named customers, assemblers and suppliers, but they are equivalent to a SC made, for example, by
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
assemblers, first tier suppliers, and second tier suppliers. The SC flexibility addressed in the paper takes into account two main aspects: (i) the process flexibility of each SC plant, concerning the number of product types that can be manufactured in each production site (supplier or assembler), and (ii) the logistics flexibility, related to the different logistics strategies which can be adopted either to release a product to a market (downstream or distribution flexibility) or to procure a component from a supplier (upstream or procurement flexibility). While the process flexibility is a well known type of manufacturing system flexibility (Garavelli, 2001), the logistics flexibility can be referred to the routing flexibility at the shop floor level, that is the ability of using alternative routes to move the work-in-process through different resources offering the same processes (Das and Nagendra, 1997; Ho and Moodie, 1996). Logistics flexibility is then here intended as the possibility of shifting the production of an item (component or final product) to different sites of a given stage of the SC, allowing to reduce the negative impact of demand and process variability on SC performance. As far as the process flexibility is concerned, the costs associated to the plant capability of producing different items usually grow with the number of items. In most cases then, due to the high flexibility investments and management costs required, it is not economic to produce all the items demanded by the global market in all the plants of a company (total flexibility). This would be the case of a multi-domestic configuration, where local plants have to deal with a certain product differentiation from a local market to another. From the opposite side, the focused factory configuration of a company, i.e. the specialization of each plant in a few products (such as a product family) not released by other plants (no flexibility configuration), can also prove neither efficient nor effective, due for instance to the higher logistics costs and lower system flexibility, respectively. An intermediate level of process flexibility has been introduced by Jordan and Graves (1995). Referring to the production planning of a multi-
143
plant company, they have shown the advantages of limited flexibility, i.e. a configuration of plants and products that, chained together forming the longest close loop, provides most of the benefits of the total flexibility configuration without requiring the correspondent investments and management costs. The concept of limited flexibility is here considered in order to analyze three possible situations for any production site (supplier or assembler): each site produces (a) one, (b) a limited number, and (c) all the product (or component) types, corresponding to the degrees of no flexibility, limited flexibility, and total flexibility, respectively. In this paper, some operational aspects not investigated in the Jordan and Graves work (Jordan and Graves, 1995) are considered, such as logistics issues, multi-stage operations networks (SC), work-in-process and time performance. The logistics aspects here considered refer to both distribution (products from assemblers to customers) and procurement (components from suppliers to assemblers). In order to consider the product distribution flows and the related logistics impact, the market demand is not considered as a whole, but it is split among the various final markets. For the sake of simplicity and generality, the market mix, that is the range of product types demanded by every market, is considered invariant. The logistics performance of a SC is also affected by the supply strategy: for instance, components can be delivered to a production plant from a local and/or from a distant supplier, as well as by single, double, or multiple sourcing. The choice of a supply strategy depends, for instance, either on the critical role of the component or on the logistics complexity (for instance, commodity parts and big components are usually provided by local suppliers). Different distribution and procurement policies are considered in the paper. In particular, each assembler can purchase the needed components from (according to an increasing value of flexibility): one (local), more than one (limited), all the available suppliers (global), respectively. Similarly, different distribution policies are considered: each assembler can deliver its products to the close market (local), to
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
144
more than one market (limited), and/or to all the markets (global). Finally, differently from the static perspective of the demand assignment problem (Jordan and Graves, 1995; Albino and Garavelli, 1999), a dynamic simulation study is carried out to measure system performance such as lead time and work-in-process for different SC configurations, characterized by variable demand rates, process times, and transport times. The SC behavior is thus analyzed by the material flows throughout the whole system, from the suppliers to the assemblers and final markets.
3. Flexibility configurations In this paper, nine configurations of the SC network are considered, resulting from the combination of the three degrees of supplier and assembler flexibility, i.e. no flexibility, limited flexibility and total flexibility, respectively. As far as the assembling phase is considered, in correspondence of the no flexibility configuration (focused factories), every plant produces just one product (Fig. 1a). In this case, every focused plant then provides every market with its own product. This involves distribution links between every plant and market. At the other extreme, in correspondence of the total flexibility configuration (multidomestic scenario), every plant manufactures all the products for its own local market (Fig. 1b). It is then reasonable that, due to the investments made, the main distribution links Products
involved are those between local plants and markets, unless unexpected events occur, requiring deliveries from distant plants. These exceptions, however, are not considered in the study. Between the two opposite flexibility extremes, there are many different system configurations. In particular, a limited flexibility configuration is here considered (Fig. 2). According to the previous definition of limited flexibility and considering a system made by N plants and N markets (each market demanding the same N products), N 2 configurations can be considered, each characterized by Q products manufactured by each plant, with 2pQoN (Fig. 2). When a higher Q is considered, the flexibility of each plant increases. However, for the sake of brevity, we consider just the limited flexibility configuration with Q ¼ 2 (Fig. 2a), since it provides the best cost–benefit trade-off, at least for a single stage of the SC (i.e. supplier–assembler or assembler–market) (Garavelli, 2001). With the limited flexibility, every plant is potentially connected with all the markets, similarly to the no flexibility configuration (Fig. 1a). Obviously, there is a particular benefit for the network having each plant deliver its two products to its local market, instead of having the two products delivered from distant plants. As far as the supplier flexibility is concerned, it is possible to extend the above considerations. In particular, considering the case of only one type of component that can be provided in N versions (one for each product), in correspondence of the no flexibility configuration every supplier plant
Plants
Markets
Products
Plants
Markets
A
P1
M1
A
P1
M1
B
P2
M2
B
P2
M2
C
P3
M3
C
P3
M3
D
P4
M4
D
P4
M4
E
P5
M5
E
P5
Product assignments
Product deliveries
(a) Focused factory configuration: no flexibility
Product assignments
M5 Product deliveries
(b) Multidomestic configuration: total flexibility
Fig. 1. Extreme flexibility configurations.
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
145
Products
Plants
Products
Plants
Products
Plants
A
P1
A
P1
A
P1
B
P2
B
P2
B
P2
C
P3
C
P3
C
P3
D
P4
D
P4
D
P4
E
P5
E
P5
E
P5
Product assignments (a) Configuration Q = 2
Product assignments
Product assignments
(b) Configuration Q = 3
(c) Configuration Q = 4
Fig. 2. Limited flexibility configurations.
Suppliers a
Assemblers A
Markets A/B/C/D/E
a/b/c/d/e
Assemblers A/B/C/D/E
Markets A/B/C/D/E
b
B
A/B/C/D/E
a/b/c/d/e
A/B/C/D/E
A/B/C/D/E
c
C
A/B/C/D/E
a/b/c/d/e
A/B/C/D/E
A/B/C/D/E
d
D
A/B/C/D/E
a/b/c/d/e
A/B/C/D/E
A/B/C/D/E
e
E
A/B/C/D/E
a/b/c/d/e
A/B/C/D/E
A/B/C/D/E
Products
Products
Components
Products
Products
(a) Configuration 1: no flexibility supply chain
Suppliers
Components
(b) Configuration 9: total flexibility supply chain
Fig. 3. Flexibility configurations with local supply.
produces just one component version, while in the total flexibility configuration every plant manufactures all the component versions required by every product type. In between, there is the limited flexibility configuration, where every plant produces two versions of the component. In this stage of the SC, since the assembler plants demand different products according to their planned flexibility, the links between suppliers and assemblers do not depend only on the supplier flexibility, but also on the assembler one. In particular, there are two situations where the local transport of the supplies is prevalent, that is when suppliers and assemblers concurrently adopt either the no flexibility (Configuration 1, Fig. 3a) or the total flexibility (Configuration 9, Fig. 3b). Despite the production similarities of suppliers
and assembler in these cases, they are still considered separately, because no vertical integration is possible and each plant formally maintains its characteristics, such as processing and transport uncertainty. There are also three situations (Configurations 2, 4 and 5) where each supplier can deliver its components only to a limited number of clients (‘‘limited’’ supply flow). This happens when either the assemblers adopt the limited flexibility configuration and the suppliers are focused (no flexibility) on one component (Fig. 4a), or vice versa (focused assemblers, limited flexibility of suppliers) (Fig. 4b), or both of them adopt the limited flexibility configuration (Fig. 4c). In all the other situations, the supplier plants can deliver their components to all the assembler plants.
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
146 Suppliers
Assemblers
Suppliers
Assemblers
Suppliers
Assemblers
a
A/B
a/b
A
a/b
A/B
b
B/C
b/c
B
b/c
B/C
c
C/D
c/d
C
c/d
C/D
d
D/E
d/e
D
d/e
D/E
e
E/A
e/a
E
e/a
E/A
Components
Products
(a) Configuration 2
Components
Products
(b) Configuration 4
Components
Products
(c) Configuration 5
Fig. 4. Flexibility configurations with limited supply flow.
In Fig. 5, a summary of the characteristics of the nine flexibility configurations is reported.
4. A simulation model A simulation model has been defined for the performance evaluation of the nine flexibility configurations. Each configuration is characterized by the same number N of suppliers, assemblers, markets, and products. This choice was aimed at providing results focused on the analysis of the different degrees of flexibility, consistently with other works in the field (e.g., Garavelli, 2001; Jordan and Graves, 1995). Each market demands all the products produced by the company. Product demands are modeled as stochastic variables exponentially distributed, with the same mean l and equal for all the markets. The production process in each plant is modeled as a queue system with a single server, infinite queue and FIFO service rule. The processing time is modeled by exponentially distributed stochastic variables, equal for every plant of a given configuration, but dependent on the system configuration. In fact, as the flexibility grows, a penalty is given to the mean production time (1=m), in order to represent the increased complexity of information and material flows, setups, etc. In particular, it is assumed that, as the flexibility increases from the no flexibility to the limited
flexibility configuration, and from the limited flexibility to the total flexibility configuration, a penalty on the mean production time is given, equal to the 10% of the mean production time of the less flexible configuration, either for supplier or for the assembler plants. This estimation, aimed at representing the increase of material flow congestion, has been set thanks to practitioners’ experience and perception (its evaluation was asked to managers of companies in the furniture and automotive industries) and it is consistent with the time penalty due to flexibility increase that, often referred to machines instead of plants, can be widely found in the literature. However, to verify the influence of this parameter on the system performance, also the case of no penalty for flexibility has been analyzed. The product/component distribution processes are modeled by assigning a transport time to each product/component. In particular, transport times are modeled by uniformly distributed stochastic variables, characterized by a shorter mean for the local transport and a longer mean for the distant transport, according to the ratio 1:4, which can be considered realistic in many situations. A heuristics has been considered for the product assignments to the supplier and assembler plants of a given SC configuration (a flow chart is reported in Fig. 6). Starting from the market demand generation, the heuristics analyses all the paths the demanded products can follow to be
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
Supplier Flexibility
No
Limited
Total
Configuration 1
Configuration 2
Configuration 3
Focused assembler and global delivery
L.F. assembler and global delivery
Flexible assembler and local delivery
Focused supplier and local supply
Focused supplier and limited supply
Focused supplier and global supply
Configuration 4
Configuration 5
Configuration 6
Focused assembler and global delivery
L.F. assembler and global delivery
Flexible assembler and local delivery
L.F. supplier and limited supply
L.F. supplier and limited supply
L.F. supplier and global supply
Configuration 7
Configuration 8
Configuration 9
Focused assembler and global delivery
L.F. assembler and global delivery
Flexible assembler and local delivery
Flexible supplier and global supply
Flexible supplier and global supply
Flexible supplier and local supply
No
Limited
147
Total
Assembler Flexibility Fig. 5. Characteristics of the nine configurations.
finally delivered to the market, according to the assembler plants that can release that product and to the supplier plants that can deliver its components. The heuristics selects the path that allows the shortest lead-time for a product to be released, taking into account the current (at the time of the demand generation) workload of the plants. The heuristics then finds the lowest expected lead time at a given time tEðLTÞt ; expressed as the sum of the expected throughput time of the supplier ms and of the assembler ma (considering components and products both in process—N—and in the queue—NQ), and of the transport times of the supply (Ts ) and of the delivery (Td ), for each supplier–assembler–market path p of a given SC configuration:
the supplier assignment), since the assembler plant assignment occurs afterwards, when the SC workload may be different. Once the product demand is assigned to a given path, this cannot be changed during the product completion. The simulation of the production network behavior has been carried out for the nine flexibility configurations of the SC. The analysis of the results refer to: *
*
*
EðLTÞtp ¼ 1=ms ðN þ NQÞ þ 1=ma ðN þ NQÞ þ Ts þ Td : This rule is addressed to select the path that minimizes the total lead-time, even if it is optimal only at the beginning of the SC (i.e. at the time of
*
an initial expected demand of: l ¼ 1=3 ¼ 0:333; then values of: l ¼ 0:357; 0:385; 0:417; expected processing times: ma ¼ ms ¼ 1=0:65; 1/ 0.715, 1/0.78, for the no flexibility, limited flexibility, and total flexibility configurations, respectively; distant transport times: TTd ¼ 2ma ¼ 2ms ; local transport times: TTl ¼ 1=2ma ¼ 1=2ms :
The main performances analyzed in the simulation are lead times, utilization rates and work in process at both the supplier and assembler plants, as well as globally.
148
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
Selection of a configuration
Definition of all the configuration paths Market demand generation Dt Selection of the paths available for the demand Dt
Computation of the Lead Time E(LT)tp for each path p
Selection of the best path: min E (LT)tp
Assignment of Dt to the path upstream plant (supplier)
Fig. 6. Flow chart of the demand assignment heuristics.
5. Simulation results The simulation has been carried out in correspondence of a SC network with N ¼ 3; similarly to the planning problem investigated in Albino and Garavelli (1999). To simplify the analysis of the simulation results, it is better to summarize that, going from configurations 1 to 3, from 4 to 6, and from 7 to 9, the assembler flexibility increases and the supplier flexibility is constant, while within the groups made of configurations 1-4-7, 2-5-8, and 3-6-9, the supplier flexibility increases from lower to higher configurations and the assembler flexibility is constant. This can also be inferred by Fig. 5. In the simulation, the utilization performance has been prevalently used to verify the steady state of the SC configurations. As expected, the utilization rates grow with both the flexibility increase, mainly due to the time penalties applied to the
plant operations, and the demand increase. However, the steady state of the system configurations was guaranteed in every workload situation. As far as the WIP and lead-time analysis are concerned, they have shown that the configurations with limited flexibility of either supplier or assembler (Configurations 2, 4, 6, 8) provide good performance, due to the balance between, from one side, the capability of reacting to the uncertainty and variability of the environment and, from the other side, the limited increase of the material flow congestion (Figs. 7 and 8). In particular, the best performance of Configuration 5, characterized by the limited flexibility of both assembler and supplier plants, emerges. It is interesting to note that the results show similar SC performance in two of the three groups of configurations characterized by constant supplier flexibility: 4-5-6 (limited supplier flexibility) and 7-8-9 (total supplier flexibility). In each of these groups, the intermediate configurations (5 and 8), characterized by the limited flexibility of the assemblers, always perform better than the other two. Then, they represent an optimal tradeoff between the absence of flexibility and the total flexibility of the assemblers. However, the first group (1-2-3), characterized by focused suppliers (no flexibility), is an exception to this rule. This means that, with no flexibility at the supplier stage, any degree of assembler flexibility is not only useless, but also detrimental, since the system performance decreases. Then, differently from the other groups, in this group the limited flexibility of the assemblers does not provide a benefit to the system, because it is not supported by any supplier flexibility. In fact, its increased capability of reacting to uncertainty does not compensate the increased congestion at the assembler plants and the distant procurement transportations (Fig. 4a). It can also be observed that, referring to Configuration 1 as a starting point (no flexibility SC), the WIP and lead time performance decreases more with the assembler plant flexibility increase (Configuration 2 or 3) than with the supplier flexibility increase (Configuration 4 or 7). Consistently, Configurations 3 and 9, characterized by the total flexibility of the assembler plants, show the worst performance. The congestion caused by
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
149
Work In Process 20 19 18 17 16 15 14 13 12
19.1 18.0 16.9
16.7
16.0
17.0
15.9
16.4
14.8
C1
C2
C3
C4
C5
C6
C7
C8
C9
Fig. 7. WIP performance (hundreds of units) of the nine configurations.
Lead Time 6.5
6.37
6
5.99 5.62
5.5
5.58
5.33
5.65
5.28
5
5.46
4.94
4.5 4 C1
C2
C3
C4
C5
C6
C7
C8
C9
Fig. 8. Lead time performance (weeks) of the nine configurations.
the total flexibility of the assembler plants (Configurations 3, 6, 9) is also confirmed by the lead times of local markets (i.e., the lead time of products whose market demand is satisfied by local assembler plants), since in such configurations, despite no product deliveries come from distant assembler plants, local lead times are the highest among all the SC configurations. These results suggest that the SC can get more benefits (or less penalties) if the SC flexibility increases at the supply stage, more than at the assembler one. This consideration recalls the bullwhip effect, that accordingly states the importance of flexibility in the upstream stages of the SC to face environment uncertainty (Lee et al., 1997). The only exception is the case of Configuration 4. In fact, starting from this configuration, characterized by limited flexibility of suppliers and
no flexibility of assemblers, it is more beneficial to increase the assembler flexibility to a limited degree (Configuration 5), than to further increase the supplier flexibility (Configuration 7). This stresses that, when the flexibility degrees of suppliers and assemblers are different, it can be more important to pursue a flexibility equilibrium among the SC stages than increasing the flexibility of a more flexible stage, even if this is the supplier one. By considering the waiting times in the plant queues it is also possible analyzing the WIP performance at the supplier stage or at the assembler stage, respectively (Fig. 9). By this analysis, it emerges how the best performance is confirmed for Configurations 5 and 4, and that the configurations with limited flexibility of the suppliers (C4, C5, C6) provide the shortest waiting
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
150
Suppliers
Queue Time
Assemblers 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 C1
C2
C3
C4
C5
C6
C7
C8
C9
Fig. 9. Waiting time in queue (weeks) for the nine configurations.
times. This result points out that the configurations with limited flexibility of the suppliers (C4, C5, C6) are better than the configurations with the same assembler flexibility and different degrees of supplier flexibility: C4 is better than C1 and C7, C5 is better than C2 and C8, C6 is better than C3 and C9 (Figs. 7 and 8). This points out that, if the degree of assembler flexibility is given, the limited flexibility of suppliers provides the best SC performance. As emerged by the simulation results, it is clear that the penalty time for flexibility has the effect of reducing the advantages of flexibility. However, a simulation has been carried out with no penalty time. This simulation, although emphasizing the benefits of flexibility, has shown that many configurations with limited flexibility (in particular, Configuration 5, but also 6 and 8) show very good performance, close to Configuration 9 (total flexibility of the SC). From the analysis of the network performance, it is also interesting to note that the configurations that show a strong difference between the degree of flexibility of assembler and supplier plants, like Configurations 3 and 7, do not behave well, in terms of both lead time and WIP. This is confirmed also from the simulation with no penalty time for flexibility. As a consequence, it can be inferred that, at any plant level, whatever investment in flexibility, which does not necessarily yield operations management benefits, should take into consideration the upstream plant flexibility opportunities. In fact, a partial optimization of a SC stage may not
produce the desired global benefits for the network. A typical example comes from the analysis of the Configurations 1-2-3, which show increasing degrees of assembler flexibility that yield SC performance decrease. The simulation has also been carried out in conditions of mean market demand increase. As expected, these situations cause a consistent decrease of system performance, in terms of WIP and lead times. As cited upwards, the utilization grows, but the system still performs in the steady state. The WIP performance, in particular, shows high levels of stock for the Configurations 3, 6, and 9 (total flexibility for assembler plants), and in particular for the last one, when the congestion of material flows affects also the suppliers (total flexibility), especially for the highest value of the demand (Fig. 10). In this case (highest demand), also the last three configurations (7, 8, 9) appear in trouble, due to the congestion of the suppliers (total flexibility) affecting every flexibility degree of the assemblers. In all these situations, the best performance is still provided by Configuration 5, the limited flexibility SC, except in the highest demand rate, when other less flexible configurations perform better (i.e., Configurations 1 and 4). Among the best four configurations, namely C1, C2, C4, and C5, the Configuration 5 is then the best for three of the four demand values, showing the best trade-off between flexibility advantages and material flow complexity.
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
151
Fig. 10. WIP (hundreds of units) of the nine configurations subject to demand increase.
The same considerations can be extended to the lead time performance. An interesting comment to this performance comes from the analysis of local and distant markets, that shows how, as the flexibility increases, the difference between local and distant market lead times tends to be negligible, for each configuration, and in particular for Configurations 6 and 8. In fact, as the plants reach the highest utilization values, the WIP tends to rapidly increase, so that the logistics influence becomes less important.
6. Conclusions In the paper, a framework for the analysis of the supply chain (SC) flexibility has been proposed. A simulation model for the evaluation of the effects of different degrees of flexibility on the performance of a SC characterized by several markets, plants and suppliers has been provided. The types of flexibility considered refer to process and logistics flexibility, defined by analogy with traditional manufacturing systems. Two extreme degrees of flexibility, namely total flexibility and no flexibility, plus an intermediate degree of flexibility, limited flexibility, have been defined. These degrees of flexibility refer to the possibility of processing a product or component in one, two, or all the SC plants, respectively. The three process flexibility degrees for each plant of the SC define nine possible configurations, with logistics connections depending on the plant
flexibility as well as on the procurement and distribution policies. The analysis of the SC configurations has been referred to the case of three products and one different component for each product. The results quantify the performance of the nine configurations considering demand variability and plant reliability. The results have shown that SC configurations with limited flexibility of either suppliers or assemblers provide good performance, due to the trade-off between the capability of reacting to uncertainty and the limited increase of the material flow congestion. In particular, the SC characterized by the limited flexibility of both assemblers and suppliers is outperforming in most cases. Only when very high demand peaks occur, configurations with focused plants perform better. The analysis has also stressed how the consistency between the supply and distribution policies (in terms of similar plant flexibility degrees) can often yield better company performance than implementing more flexibility only in one stage of the SC (assemblers or suppliers). Moreover, when this symmetry is provided, the SC gets more benefits when the SC flexibility increases at the upstream stage, more than at the downstream one, consistently with the bullwhip effect. The study has then shown how the concept of limited flexibility can be usefully applied also to the supply chain configuration. This way of
152
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153
defining suitable flexibility degrees for the different stages and nodes of the SC network appears very effective. Future developments of the study will be addressed to the consideration of more complex supply chains, in order to represent other practical implementations and results of the model.
References Albino, V., Garavelli, A.C., 1999. Limited flexibility in cellular manufacturing systems: a simulation study. International Journal of Production Economics 60–61, 447–455. Albino, V., Garavelli, A.C., Giannoccaro, I., 2002. Process and logistics flexibility for the supply chain management. Proceedings of the Seventh International Symposium on Logistics and Fifth International Symposium on Operations Strategy, Melbourne, Australia, 14–17 July. Bhatnagar, R., Chandra, P., Goyal, S.K., 1993. Models for multiplant coordination. European Journal of Operational Research 67, 141–160. Browne, J., Dubois, D., Rathmill, K., Sethi, S.P., Stecke, K.E., 1984. Classification of flexible manufacturing systems. The FMS Magazine 2 (2), 114–117. Chambers, S., 1992. Flexibility in the context of manufacturing strategy. In: Voss, C.A. (Ed.), Manufacturing Strategy— Process and Content. Chapman & Hall, London. Chen, C.-F., Egbelu, P.J., Wu, C.T., 1994. Production planning models for a central factory with multiple satellite factories. International Journal of Production Research 32 (6), 1431–1450. Christopher, M., 1992. Logistic and Supply Chain Management. Pitman, London. Chung, C.H., Chen, I.J., 1990. Managing flexibility of flexible manufacturing systems for competitive edge. In: Liberatore, M.J. (Ed.), The Selection and Evaluation of Advanced Manufacturing Technologies. Springer, Berlin. Coyne, K.P., Dye, R., 1998. The competitive dynamics of network-based business. Harvard Business Review, January–February. D’Souza, D., Williams, D., 2000. Toward a taxonomy of manufacturing flexibility dimensions. Journal of Operations Management 18 (5), 577–593. Das, S.K., Nagendra, P., 1997. Selection of routes in a flexible manufacturing facility. International Journal of Production Economics 48, 237–247. de Groote, X., 1994. The flexibility of production processes: A general framework. Management Science 40 (7), 933–945. Duclos, L.K., Lummus, R.R., Vokurka, R.J., 2001. A conceptual model of supply chain flexibility, Proceedings of the Paper Presented at the 32th Decision Science Institute Annual Meeting, San Francisco, USA, November 17–20.
Egbelu, P.J., 1991. Batch production time in a multi-stage system with material handling consideration. International Journal of Production Research 29 (4), 695–712. Garavelli, A.C., 2001. Performance analysis of a batch production system with limited flexibility. International Journal of Production Economics 69 (1), 39–48. Gerwin, D., 1993. Manufacturing flexibility: A strategic perspective. Management Science 39 (4), 395–410. Giannoccaro, I., Pontrandolfo, P., 2001. Models for Supply Chain Management: A Taxonomy. Proceedings of the POM-2001 Conference: POM Mastery in the New Millennium, Orlando, FL, March 30–April 2. Giannoccaro, I., Pontrantrandolfo, P., Scozzi, B., 2003. Uncertainty in supply chain inventory management: a fuzzy approach. European Journal of Operational Research 149, 185–196. Gupta, D., 1993. On measurement and valuation of manufacturing flexibility. International Journal of Production Research 31 (12), 2947–2958. Gupta, D., Buzacott, J.A., 1996. A ‘‘goodness test’’ for operational measures of manufacturing flexibility. International Journal of Manufacturing Systems 8, 233–245. Gupta, Y.P., Somers, T.M., 1992. The measurement of manufacturing flexibility. European Journal of Operational Research 60, 166–182. Ho, Y.C., Moodie, C.L., 1996. Solving cell formation problems in a manufacturing environment with flexible processing and routing capabilities. International Journal of Production Research 34 (10), 2901–2923. Hyun, J.-H., Ahn, B.H., 1992. A unifying framework for manufacturing flexibility. Manufacturing review 5 (4), 251–260. Jordan, W.J., Graves, S.C., 1995. Principles on the benefits of manufacturing process flexibility. Management Science 41 (4), 577–594. Kim, C., 1991. Issues on manufacturing flexibility. International Journal of Production Research 2 (2), 4–13. Kogut, B., Kulatilaka, N., 1994. Operating flexibility, global manufacturing, and the option value of a multinational network. Management Science 40 (1), 123–139. Koste, L., Malhotra, M.K., 1999. A theoretical framework for the dimensions of manufacturing flexibility. Journal of Operations Management 18 (1), 75–93. Lee, H.L., Padmanabhan, V., Seungjin Whang, 1997. The Bullwhip effect in supply chains. Sloan Management Review 38 (Spring), 93–102. Lynch, R.L., Cross, K.F., 1991. Measure up!. Blackwell, Cambridge, MA. Nakane, J., Hall, R.W., 1991. Holonic manufacturing: Flexibility—the competitive battle in the 1990s. Production Planning and Control 2 (1), 2–13. Sarker, B.R., Krishnamurthy, S., Kuthethur, S.G., 1994. A survey and critical review of flexibility measures in manufacturing systems. Production Planning and Control 5 (6), 512–523. Sethi, A.K., Sethi, S.P., 1990. Flexibility in manufacturing: A survey. International Journal of Flexible Manufacturing 2 (4), 289–328.
A.C. Garavelli / Int. J. Production Economics 85 (2003) 141–153 Slack, N., 1987. The flexibility of manufacturing systems. International Journal of Operations and Production Management 7 (4), 35–45. Upton, D.M., 1994. The management of manufacturing flexibility. California Management Review 36 (Winter), 72–89. Vickery, S., Calantone, R., Droge, C., 1999. Supply chain flexibility: An empirical study. The Journal of Supply Chain Management 35 (August), 16–24.
153
Viswanadham, N., Srinivasa Raghavan, N.R., 1997. Flexibility in manufacturing enterprises. Sadhana 22 (2), 135–163. Vokurka, R.J., O’Leary-Kelly, S., 2000. A review of manufacturing flexibility empirical research. Journal of Operations Management 18 (4), 485–501. Zelenovich, D.M., 1982. Flexibility: A condition for effective production systems. International Journal of Production Research 20 (3), 319–337.