Facilities location and layout are both inherently prone to hierarchical aggregation so as to best direct design attention and harness the complexity and scale of the design space. Figure 5.1 provides an illustra-tion of hierarchical aggregaillustra-tion. The entire network of facilities of an enterprise is depicted on the top portion of Figure 5.1, as currently located around the world. The company produces a core module in Scandinavia. This core module is fed to three regional product assembly plants, respectively located in the United States, Eastern Europe, and Japan. Each of these assemblers feeds a set of market-dedicated distribution centers. The middle of Figure 5.1 depicts the site of the Eastern European Assembler, located on municipal lot 62-32. The plan distinguishes seven types of zones in the site. Facility zones are segre-gated into three types: administration, factory, and laboratory. Transportation zones are split into two types: road zone and parking and transit zone. There is a green zone for trees, grassy areas and gardens.
Finally, there is an expansion zone for further expanding activities in the future. There are two factories on the site. The lower portion of Figure 5.1 depicts the assembly factory F2, itself comprised of a number of assembly, production, and distribution centers, as well as offices, meeting rooms, laboratories, and personal care rooms.
A modular approach to represent facility networks helps navigate through various levels of a hierarchi-cal organization. In Figure 5.1, the framework introduced by Montreuil (2006) has been used. It represents the facilities and centers through their main role in the network: assembler, distributor, fulfiller, producer, processor, transporter, as well as a number of more specific roles. A producer fabricates products, modules and parts through operations on materials. A processor performs operations on clients’ products and parts. An assembler makes products and modules by assembling them from parts and modules provided by suppliers. A fulfiller fulfils and customizes client orders from products and modules. A distributor stores, prepares and ships products, modules and parts to satisfy client orders. A transporter moves, transports, and handles objects between centers according to client orders. Montreuil (2006) describes thoroughly each type of role and its design issues. Using the same terms at various levels helps the engineer comprehend more readily the nature of the network and its constituents, and leverage this knowledge into developing better designs.
Depending on the scope of design decisions to be taken, the engineer selects the appropriate level of aggregation. However, he must always take advantage of in-depth knowledge of higher and lower levels of aggregation to leverage potential options, taking advantage of installed assets and fostering synergies.
The illustration has focused on hierarchical aggregation. In location and layout studies another type of aggregation is of foremost importance: physical aggregation. This is introduced here through a layout illustration, yet the logic is similar for multi-facility location. The layout of a facility can be represented with various degrees of physical aggregation for design purposes.
Lot 62-31 Lot 62-32
Lot 63-01
F1 F2
A L
G
G
G R
R P
P
P
E
A: administration E: expansion zone F: Assembly factory G: green zone L: laboratory P: parking & transit R: road
Module assemblers Parts producers
Parts distributor Materials & parts distributor
Product assemblers
Product distributor
Purchased modules and parts distributor
Offices, meeting rooms, laboratories and personal care rooms
Inbound
Inbound Outbound Parking
East European assembler site
Factory: F2 assembler
Green zone Global facilities network
Core module producer Regional
product assembler Market distributors
FIGuRE 5�1 Hierarchical illustration of facilities network deployment, site layout, and facility layout.
The final deliverable is to be an implemented and operational physical facility laid out according to the design team specifications. The final form of these representations is an engineering drawing and/or a 3D rendering of the facility, with detailed location of all structural elements, infrastructures, walls, machines, etc., identifying the various centers sharing the overall space. For most of the design process, such levels of details are usually not necessary and are cumbersome to manipulate.
Figure 5.2 exhibits five levels of layout representation used for design purposes. The least aggregate first level, here termed processor layout, shows the location and shape of the building, each center, each aisle and each significant processor within each center (e.g., Warnecke and Dangelmaier 1982). The processor layout also locates the input and/or output stations of each center, the travel lane directions for each aisle and, when appropriate, the main material handling systems such as conveyor systems and cranes.
At the second level of aggregation lies the net layout which does not show the processors within each center (e.g., Montreuil 1991, Wu and Appleton 2002). The assumption when focusing the design process on the net layout is that prior to developing the entire layout for the facility, space estimates have been made for each center, leading to area and shape specifications, and that as long as these spatial specifica-tions are satisfied, then the net layout embeds most of the critical design issues. The space estimation may involve designing a priori potential alternative processor layouts for each center. The transposition of the net layout to a processor layout for the overall facility is left as a detailed exercise where the layout of each center is developed given the shape and location decided through the net layout. Note that when the inter-nal layout of the centers has influence on overall flow and physical feasibility, then basing the core of the design process on the net layout is not adequate.
At the third level of aggregation, the aisle set is not included anymore in the layout (e.g., Montreuil 1987, 1991). Instead, the space requirements for shaping each center are augmented by the amount of space expected to be used by aisles in the overall layout. For example, if by experience, roughly 15% of the overall space is occupied by aisles in layouts for the kind of facility to be designed, then the space require-ments of each center are increased by 15%. This percentage is iteratively adjusted as needed. The layout depicting the location and shape of the centers is now termed a block layout.
At this third level, instead of including the aisle set explicitly, the design depicts the logical travel network (Chhajed et al. 1992). This network, or combination of networks, connects the I/O stations of the centers as well as the facility entry and exit locations. There may be a network representing aisle travel, or even more specifically people travel or vehicle travel. Other networks may represent travel along an over-head conveyor or a monorail. The network is superimposed on the block layout, allowing the easy altera-tion of one or the other without having to always maintain integrity between them during the design process, which eases the editing process. Links of the network can be drawn proportional to their expected traffic. When transposition of a block layout with a travel network into a net layout or a processor layout proves cumbersome due to the need for major adjustments, then such a level of aggregation may not be appropriate for design purposes.
At the fourth level of aggregation, the travel network is not depicted, leaving only the block layout and I/O stations (e.g., Montreuil and Ratliff 1988a). Editing such a block layout with only input/output stations depicted is easy with most current drawing packages. These stations clearly depict where flow is to enter and exit each center in the layout. Even though the I/O stations of each center can be located anywhere within the center, in practice most of the times they are located either at center periphery or at its centroid.
The former is usually in concordance with prior space specifications. It is commonly used when it is known that the center is to be an assembly line, a U-shape cell, a major piece of equipment with clear input and output locations, a walled zone with access doors, etc. The latter centroid location, right in the middle of the center, is mostly used when the center is composed of a set of processors and flow can go directly to and from any of them from or to the outside of the center. It is basically equivalent to saying that one has no idea how flow is to occur in the center or that flow is to be uniformly distributed through the center.
The absence of travel network representation assumes that the design of the network and the aisle set can be straightforwardly realized afterward without distorting the essence of the network, and that flow travel can be easily approximated without explicit specification of the travel network. Normally, one of the
two following assumptions justifies flow approximation. The first is that a free flow movement is represen-tative, computed either through the rectilinear or Euclidean distance between the I/O stations between which a flow is expected to occur. Figure 5.3 illustrates these two types of free flow. Euclidean distance assumes that one can travel almost directly from one station to another while rectilinear distance assumes orthogonal staircase travel along the X and Y axes, like through a typical aisle set when one does not have
FIGuRE 5�2 Degrees of aggregation in layout representation for design purposes.
A F
G
B E
(e) Block layout
A
F
G
B E
(d) Block layout with I/O stations
A
F
G
(c) Block layout with travel network
A
C F
G B
(b) Net layout
F E B
D A
C
(a) Processor layout
E
B E
D
C D
C D
C D
to backtrack along any of the axes. The second alternative assumption is that flow travel is to occur along the center boundaries. Thus distances can be measured accordingly through the shortest path between the two I/O stations of each flow, along the contour network of the facility. This network is implicitly created by inserting a node at each corner of one center and/or the facility, and inserting a link along each center or facility boundary segment between the nodes. In Figure 5.2d, a flow from the northern output station of center B to the input station of center G would be assumed to travel from the output station of B southward along the west boundary of center B, then turning eastbound and traveling along the southern boundaries of center E, and keeping straight forward to reach the input station of center G.
At the fifth level of aggregation, only the block layout is drawn in Figure 5.2e. This is the simplest representation. On the one hand it is the easiest to draw and edit. On the other hand it is the most approxi-mate in terms of location, shape, and flow. For the last 50 years, it has been by far the most commonly taught representation in academic books and classes, often the only one (e.g., Tompkins et al. 2003), and it has been the most researched. It is equivalent to the fourth level with all the I/O stations located at the centroid of their center. The underlying assumption justifying this level of aggregation is that the relative positioning and shaping of the centers embeds most of the design value and that this positioning and shaping can be done disregarding I/O stations, travel networks, aisle sets and processors, which are minor issues and will be dealt with at later stages. While in some settings this is appropriate, in many others such an aggregation can be dangerous. It may lead to the incorrect perception that the implemented layout is optimal because its underlying block layout was evaluated optimal at the highest level of aggregation, thus limiting and biasing the creative space of designers.
It is always a worthwhile exercise, when analyzing an existing facility, to draw and study it at various levels of aggregation. Each level may reveal insights unreachable at other levels, either because they do not show the appropriate information or because it is hidden in too much detail.