309

Invited Review

Operational research and material handling *

Jessica O. M A T S O N a n d J o h n A. W H I T E changes. Not many years ago, it seemed sufficient

School of lndustrtal and Systems Engineering, Georgta lnstttute to define material handling as simply handling
of .,echnology, Atlanta, GA 30332, U.S A. material. Today, material handling is defined much
more broadly. As an example, one might define
R._~,eived March 1982
material handling as using the right method to
provide safely the right amount of the right matertal
without damage at the right place, at the rtght ttme,
A number of material handling research areas are reviewed
and opportunities for further research are identified. Included in the right condition, and at the right cost.
in the review is a consideration of the following areas: robotics, Some continue to view material handhng as
ce,.weyor theory, transfer lines, flexible manufacturing systems, "lift that barge and tote that bale". Others view
eqmpment selection, storage altervatives, automated storage material handling to be handling, storing, and
and retrieval systems, warehouse layout, palletlzmg, and order
controlling material, with considerable emphasis
picking and accumulation.
on the latter.
The Factory of the Future, also known as the
I. Introduction automatic factory, has captured the attention of
upper management in corporations throughout the
In recent years there has developed a body of world. The recent announcements from Japan's
operational research literature that could be loosely Fujitsu Fanuc and Yamazaki Machinery Works of
classified as material handling research. Because of unmanned factory systems, coupled with the broad
the growing interest in both the design of new or application in Japan of quality circles, rapid tool
improved material handling systems and the oper- setting, and "The Toyota Production System". have
ation of existing systems, it seems appropriate and generated intense pressures among the developed
timely to provide an overview of the field and to countries to reindustrialize and t~ d e s i ~ auto-
focus on future research needs. Design issues typi- mated factory systems.
cally are focused on strategic aspects of the sys- The emphasis on automated factory systen~,, has
tem; whereas, operational issues are generally focused considerable attention on the use of auto-
focused on tactical aspects. mated material handling systems. Unfortunately,
The increasing capability to utilize the com- the rush by industry to install automated equip-
puter to control material handling systems has ment has resulted in the creation of islands of
resulted in a surge of interest in the general field of automation, rather than integrated material han-
material handfing. The microprocessor revolution dhng systems.
has certainly made its mark on the field. The role Typical examples of islands of automation in-
and scope of material handling, as well as material clude numerically controlled machine tools, robots,
handling technology, have undergone dramatic automated storage/retrieval systems, automated
guided vehicles, automatic palletizers, automatic
* The preparation of thts review was supported by the office of
N~,val Research under Contract No. N00014.80-k-0709, and tool changers, transfer lines, and flexible manufac-
the National Science Foundation under Grant No. 1SI- turing systems. In some cases, the islands are the
81 ! 5823. size of individual work ztations or machine centers;
in other cases, the islands are the size of manufac-
North-Holland Publishing Company turing departments.
European Journal of Operational Research 11 (1982) 309-318 From a systems viewpoint, islands of automa-

037%2217/82/0000-0000/$02.75 © 1982 North-Holland

 J.O. Matson, J.A. White / Material handhng
310

tion are not necessarily bad, so long as they are some aspect of material handling. Only a brief
considered to be interim objectives in a phased amount of attention is given to the application of
implementation of an automated system. How- simulation in modeling material handling systems.
ever, to obtain an integrated factory system, the The imbalance in our coverage of simulation is
islands of automation must be tied together or due to our belief that more research attention is
linked. An obvious approach that can be used to needed in the development of analytical models.
physically bridge the islands of automation is the There have been many very successful applications
material handling system. of simulation in designing material handling sys-
Ver ~ittle basic research has been devoted to tems, Phillips [461, Rearick et al. [49], and Swain
issues related to the design of integrated material and Marsh [59].
handling s'~stems. Indeed, few have addressed the
subject of material handling at all. Those who
have studied explicitly material handling problems 2. Robotics
have tended to focus on a very narrowly defined
aspect of the subject; counter-examples would in- Robotic research has been directed toward either
clude large scale simulation efforts. the design of robot hardware a n d / o r software
Even though there has been a limited amount of systems or the integration of a robot in a manufac-
research performed that addresses explicitly turing or assembly environment. As examples,
material handling problems, mucl~ of the research Medeiros et al. [39] developed a simulation model
devoted to other, more generic, problems is appli- for a mare+facturing cell that included a robot with
cable to material handling. Some examples of re- vision; the production setting was assumed to be
search areas that realistically contribute to a better low-volume, batch manufacturing. A detailed
understanding of material handling problems in- simulation model was developed by Stold~ and
clude ergonomics, man-machine systems, vehicle Vukobratovi/: [58] to study the insertion of a peg
routing, facilities layout and location, scheduling into a hole; the model included the consideration
and sequencing, queuing theory, and computer of manipulator dynamics during the assembly pro-
storage systems design, among others. A brief de- cess; their analysis included trajectory control and
scription of some of the generic research areas of feedback in the case of interference. Nof et al. [44]
operational research that relate to material han- studied the comparative abilities and limitations of
dling is given by Maxwell [37]. industrial robots and humans; they developed a
In this review paper, the coverage of opera- robot oriented job and skills analysis method to
tional research applied to material handling is optimize the task performance for the robot and to
limited to the following ten categories: robotics, aid in the selection of specific robots for specific
conveyor theory, transfer lines, flexible manufac- tasks. Vereschagin and Generozov [61] used dy-
turing systems, equipment selection, storage alter- namic programmiag to determine the optimal
natives, automated storage and retrieval systems, trajectory of the robot arm in performing a given
warehouse layout, palletizing, and order picking task.
and accumulation. It is emphasized that the set of
ten categories does not represent a definition of
material handling; the set is neither exhaustive nor 3. Conveyor theory
m~ually exclusive. Furthermore, the coverage of
any specific category is not intended to be a In comparison with the other research cate-
comprehensive survey of the literature. gories, there exists a large literature devoted to
The objective of the review is to acquaint the conveyor theory. Muth and White [43] reviewed
reader with the breadth and depth of the research the conveyor theory literature through 19"/7; de-
that has been devoted to material handling. Addi- terministic and probabilistic models of conveyors
tionaUy, it is hoped that the review will stimulate were considered, as well as descriptive and norma-
additional research on material handling prob- tive approaches to modeling conveyor systems.
lems. The review focused on the study of operational
Mo~t of :he research described in the review characteristics of conveyor systems. The state-of-
deals with the development of analytical models of the-art for conveyor theory research is perhaps
 J.O. Matson, J.A. White / Material handhng 311

best represented by the contributions of Muth (a) Constant,

[41,421, Gregory and Litton [20], and Maxwell and (b) Random;
Wilson [38]. (2) Machine failures:
Muth [41] developed a deterministic model of (a) Do not occur,
material flow on a conveyor having multiple load- (b) Occur deterministically,
ing and unloading stations, equally spaced car- (c) Occur randomly:
riers, and periodically time-varying patterns of (i) Independently,
material flow; the stations can perform either pure (ii) Dependently;
loading, pure unloading, or alternating loading (3) Machine downtime durations:
and unloading operations. Material flow along the (a) Zero,
conveyor is described by a difference equation (b) Constant,
whose solution yields steady state conditions for (c) Random;
general flow patterns. (4) Scrapping of material being worked on when
In [42] Muth considered random material flow a machine fails:
for a conveyor with discretely spaced carriers and (a) None scrapped,
having a single loading station and a single un- (b) All scrapped,
loading station. He found that the conveyor could (c) Randomly scrapped;
serve as a variance-reducer; however, to decrease (5) Buffer storage capacities:
the output variance required an increase in con- (a) Zero,
veyor capacity. (b) Finite,
Gregory and Litton [20] modeled probabilisti- (c) Infinite;
tally a conveyor with discretely spaced carriers (6) Inter-arrival times at first machine:
serving multiple unloading stations; no recircula- (a) Zero,
tion of material was allowed. All carriers ap- (b) Random;
proaching the first station were assumed to be (7) Material flow:
exponentially distributed; if an unloading attempt (a) D~screte
is unsuccessful due to a carrier being empty, the (b) Continuous;
operator waits for the next loaded carrier to arrive. (8) Method of modeling:
(Subsequently, Ohta [45] developed a GERT net- (a) Analytical,
work representation of the conveyor system studied (b) Simulation.
by Gregory and Litton; Ohta's formulation allows
the use of general probability distributions.) Relatively few of the possible combinations of
Maxwell and Wilson [38] developed a network the eight categories given above have been ad-
model of a conveyor system; they modeled the dressed. In general, the research performed on
dynamic performance of the conveyor system by transfer lines having random components has as-
using a time expanded network flow model. Mod- sumed memoryless distributions or combinations
els were developed for a variety of conveyors and of memoryless distributions.
their results were compared with industrial prac- The publications by Buzacott and Hanifin [ 10],
tice. Gershwin and Berman [15], and lgnall and Silver
[26] are represetitative of the contributions to the
research on transfer lines. Buzacott and Hanifin
4. Transfer lines [10] evaluate and compare seven analytical models
of transfer lines; they also compare the model
A transfer line consists of several machines results with real data from a transfer line. Gersh-
linked together in series by a conveyor; buffer win and Berman [15] consider a transfer line for
storage is frequently provided between the ma- discrete parts consisting of two machines and a
chines. Transfer lines are typically modeled as a single finite buffer; exponential service, failure,
series of queues. and repair processes are assumed; they also pro-
Research on transfer lines can be classified vide an extensive listing of research papers treating
using the following criteria: analytical models of transfer lines. Ignall and Silver
(1) Processing times: [26] consider a generalization of the transfer line
312 J O. Matson. J.A. Whtte / Material handhng

that can be represented as a series of two multi- 6. Equipment selection

server queues: a two-stage system is considered,
with multiple machines available at each stage: Several heuristic procedures have been devel-
they estimate the output from the system as a oped for selecting the appropriate materials han-
function of the size of the inter-stage buffer. dling equipment from a large number of alterna-
tives. Jones [27], for example, formulated the
equipment selection problem as a network flow
5. Flexible manufacturing systems problem in which the nodes represented equip-
ment alternatives for each warehousing function
Where transfer lines can be modeled as a series and the arcs represented the compatibility of
of work stations, a flexible manufacturing system equipment alternatives serving successive ware-
can be represented as a network of work stations. house functions. His assumptions included specific
The flexibdity is obtained from the use of general equipment for each warehouse function, no mixed
purpose machines and a material handling system mode equipment selections for warehouse func-
that allows multiple routings of material through tions other than retrieval of goods from storage,
the machines..Typically, the flexible manufactur- no cost for goods' processing and waiting times,
ing system is under central computer control. arrival and storage of goods in unitized form, and
(Hutchinson and Wynne [25] provide a detailed retrieval of goods by individual cartons. The solu-
description of flexible manufacturing systems.) tion procedure required determining the least-cost
With a transfer line, all parts follow the same paths through separate receiving and shipping net-
sequence of operations. When a machine fails, the works. Simulation was used for queuing investi-
transfer hne is generally blocked. With a flexible gations to develop approximations for parameters
manufacturing system, different material routings associated with throughput requirements.
are achieved either by "providing separate flow Assuming a given set of moves, Webster and
paths between each pair of machines" or by "using Reed [62] formulated the problem as an assign-
a common material handling device through which ment problem and developed a solution procedure
all parts pass and which connects all machines". based on specification of a cost matrix for each
[ll]. move-equipment combination. The procedure con-
Buzacott and Shanthikumar [ l l] developed ana- sisted of two main steps:
lytical models of a number of flexible manufactur- (!) determination of an initial feasible solution,
ing systems: they also considered a number of and
alternate storage and control alternatives. Based (2) improvement of the initial solution by con-
on their research, it was concluded that the ef- sidering system utilization.
fectiveness of the flexible manufacturing system is Combining the heuristic equipment selection pro-
strongly dependent on the effectiveness of the cedure of Webster and Reed with the CRAFT
control system. facilities layout procedure [1,9], Tompkins and
In addition to the analytical models studied by Reed [60] developed the COFAD (COmputerized
Buzacott and Shanthikumar [ll], a number of FAcilities Design) model COFAD jointly selects
simulation models of flexible manufacturing sys- the facility layout and materials handling equip-
tems have been developed. Hughes [24] describes a ment combination which minimizes material han-
simulation model that is used by a flexible manu- dling costs.
facturing system supplier in designing such sys- Graves and Whitney [19] examined the equip-
tems. Biles and Yang [7] also developed a simula- ment (resource) selection and assignment problem
tion model of flexible manufacturing systems: they in the context of assembly systems design. For a
focused on the material handling componeats, with mixed-integer linear programming formulation of
specific attention given to the use of multiple the problem, they proposed a branch and bound
parallel conveyors. solution procedure in which bounds were de-
termined by subgradient optimization. The solu-
tion procedure 'was demonstrated for the design of
a robot assembly system. Graves and Lamar [18]
extended the work of [19] and established a more
 J.O. Matson, J.A. WhJte /Matertal handhng 313

general integer programming formulation. An ap- of determining the optimum lane depths for multi-
proximate solution procedure which provided in- ple products when there is a constraint on the
teger assignments was given. number of depths allowed.
Using simulation, Marsh [34] compared three
alternative policies for the operation of block
7. Storage alternatives stacking storage systems. He examined the effect
on space utilization of storing in longer or shorter
Examples of unit load storage alternatives ili- lanes when the optimum lane depth is not availa-
clude block stacking, pallet racks, and deep lane ble.
storage systems. Much of the research on the
design and evaluation of these storage alternatives
has been concentrated on the criterion of space 8. Automated storage/retrieval systems
utilization. Moder and Thornton [40], for example,
developed quantitative models for block stacking The design and operation of automated storage
to determine the effect of clearance between lanes and retrieval systems have been studied by a num-
and the angle of the storage lanes on space utiliza- ber of researchers. Hausman, Schwartz and Graves
tion, assuming maximum occupancy of the storage [21], Graves, Hausman and Schwarz [17], and
positions. Kind [30,31] demonstrated that space Schwarz, Graves and Hausman [53] studied the
utilization should be considered, not only at peak assignment of multiple items to the same pallet,
inventory, but over the entire inventory cycle of a the assignment of pallet loads to storage locations.
product; he proposed a formula for block stacking and the sequencing of storage and retrieval re-
lane depth which approximately maximized space quests. They considered both dedicated (fixed-slot)
utilization for a product whose inventory cycle was storage and randomized (floating-slot) storage.
characterized by uniform withdrawal of stock, in- White and Bozer [63] extended the research of
stantaneous replenishment, and zero safety stock. Graves, Hausman and Schwarz to include eco-
Kind's study [30] was the first of several to nomic considerations. Using cost estimates devel-
consider the design of storage systems for maxi- oped by Zollinger [66], an optimization model was
mum space utilization over a product's inventory developed to determine the number, length, and
cycle, given the assumptions of randomized stor- height of storage aisles to minimize the after-tax,
age, FIFO lot rotation, and unit load storages and present worth cost of the combination of storage
retrievals. Kooy [32], Kooy and Peterson [33], and racks, storage machines, building, and land.
DeMars [12] reported on the solution of some Karasawa, Nakayama alld Dohi [29] developed
actual storage desi~n problems in industry. White, a non-linear mixed-integer programming formtda-
DeMars, and Matson [64] developed analytical tion of an AS/RS. Their cost model also included
models to determine the minimum space designs consideration of storage racks, storage machines,
for block stacking, single-deep and double-deep building, and land. However, their model differed
pallet rack, deep lane storage, and unit load significantly from that developed by Zoilinger [66].
AS/RS. Three storage-retrieval distributions were Bozer, Shieh and White [8] developed a proce-
considered including a uniform withdrawal rate, dure for determining travel time for both single
an increasing withdrawal rate, and a decreasing command and dual command cvcles for an AS/RS
withdrawal rate. Extending the analysis of [64], when horizontal and verucal travel occur simulta-
Matson and White [35] demonstrated the effect of neously and the storage/retrieval horizontal and
handling requirements on the optimum storage vertical addresses are independently distributed
design and developed a total cost model for block random variables. Additionally, they developed an
stacking incorporating both space and throughput interactive procedure for determining the dedi-
costs. In addition, block stacking models for prod- cated storage assignment of pallets to storage slots
ucts with bulk withdrawals were given. Matson, to maximize throughput under single command
Shieh and White [36] used a dynamic program- and dual command conditions; a colorgraphics
ming approach to determine the minimum space computer was used to display layout information
block stacking lane depths for the case of lot to the user.
splitting. In addition, they considered the problem A number of analytical and simulation models
314 J.O. Matson, J A. White /Matertal handhng

have been developed by AS/RS equipment sup- 10. Palletizing

pliers and user companies. Much of the work has
focused on specific applications and design ques- Palletization research has concentrated on three
tions. Examples of such work include the efforts of topics:
Barrett [3] and Sand [5i]. (1) setting the optimum pallet size,
(2) selecting the optimum pallet layout, and
(3) scheduling the items to be palletized.
9. Warehouse layout Steudel [56] considered the effect of pallet size on
materials handling, storage system, and floor space
The location of inventory items in the warehouse costs. To determine the optimum pallet size, he
to minimize material handling time or costs is one developed a computer model to determine the
aspect of the warehouse layout problem which has optimum pallet layout and to evaluate a total cost
received much attention. Kallina and Lynn [28] function for each alternative pallet size.
discussed the application of the cube-per-order Steudel [57] also examined the pallet loading
index rule, first proposed by Heskett [22], for problem as a special case of the two-dimensional
assigning warehouse space to inventory items. cutting stock problem. Assuming that all rectan-
Francis and White [14] investigated both discrete gles were of equal size and allowing nonguillotine
and continuous layout problems and demon- cuts, he proposed a heuristic procedure based on
strated that the cube-per-order rule minimizes dynamic programming. The heuristic, which pro-
handling costs when a "factoring" assumption is vided a cyclic four module layout pattern, first
satisfied. A contour line solution procedure was maximized utilization of the pallet perimeter and
given for continuous layout designs. then projected the perimeter arrangement inward
Employing the cube-per-order rule, Wilson [65] toward the pallet center. Correction steps were
jointly considered the relationship of the stock required for overlapping rectangles and large
location and order quantity problems. An iterative central holes.
gradient search procedure was used to solve the Others who have proposed pallet layout heuris-
nonlinear programming formulation. Whereas tics for similar assumptions include Smith and
Wilson examined the discrete layout problem, DeCani [551 and Ratliff and Tran [48]. Smith and
Hodgson and Lowe [23] developed an algorithm to DeCani described the implementation of an enu-
find locally minimum cost solutions for the con- merative procedure to evaluate all possible cyclic
tinuous design problem. layouts. Ratliff and Tran developed an optimum
Another element of warehouse layout involves solution method which takes advantage of the
determining the actual configuration of storage mathematical properties of the four module pat-
bays, aisles, and doors to minimize handling costs. tern to reduce the number of layouts required for
Roberts and Reed [50] used dynamic program- evaluation.
ming to determine the optimum configuration of Bartholdi [5] examined the problem of sched-
storage bays to minimize handling and construc- uling a sing!e paUetizer for two different items; he
tion costs. Their assumptions included the follow- established necessary and sufficient conditions for
ing: a single dock location, randomized storage, a zero delay schedule, developed minimizing func-
FIFO inventory polky, rectilinear travel, and stor- tions of delays, and discussed problem complexity.
age bays of identica! size and orientation. Bassan, A somewhat different scheduling problem was
Roll and Rosenblatt [6] investigated the optimum studied by Santana and Platzman [52]. For a single
number of racks or ~helves, and ghe length of rack palletizer serving N accumulation lanes for N types
sections for two warehouse layout configurations, of items, they investigated the problem of de-
handling, floor space, and building perimeter costs termining the best service policy to minimize the
were included in the analysis; and some rules were penalty rate; a penalty was charged when a full
developed, based on cost ratios, regarding the pre- accumulation lane caused the rejection of an arriv-
ferred warehouse layout. ing item, and the service policy determined the
next type of item to be served.
 J.O. Matson, J.A. Whtte/ Material handhng 315

11. Order picking and accumulation An order accumulation problem studied by

Elsayed [13] involved the determintion of the opti-
The order picking problem can involve either a mum combination of items to be included in one
person or a machine traveling to each of several batch or tour. Assuming an automated storage/re-
storage localions where items are removed from trieval machine for handling, he compared four
storage. Alternately, the storage location can be heuristic approaches for grouping items into
delivered to a stationary picking station where a batches. The heuristics differed only in the method
person or machine removes the required items of selecting the first item to be included in the
from storage. batch, and tours or batches were determined by
The order accumulation problem involves a solving the traveling salesman problem. Another
sorting of the items picked into customer group- study of optimal batching was conducted by
ings. The order accumulation problem also involves Armstrong, Cook, and Saipe [2]. The batching
the unitizing of loads (palletizing) and batching of problem for a conveyorized semiautomated
items or orders. orderpicking system was formulated as a mixed-in-
An order picking problem studied by Ratliff teger linear programming problem, and Benders'
and Rosenthal [47] involves the determination of decomposition was used to solve the problem.
the sequence or route which minimizes the time or
distance required for retrieval of a group of items
(an order) from storage. Their study considered a 12. Research opportunities
rectangular warehouse and assumed that cross-
overs were allowed only at the ends of aisles. A Material handling problems, in general, are very
graph theory based algorithm was developed for difficult problems to solve. Performance and eco-
determining the optimum (minimum time) solu- nomic characteristics of the technological alterna-
tion. tives are not well understood; numerous equip-
Goetschalckx and Ratliff [161 examined the ment alternatives are available for consideration
problems of where to store items, how to sequence and the number is steadily increasing; the
orders, and how to sequence items within an order. trade-offs inherent in choices between storage and
They discussed solution procedures which are ap- throughput are not defined; and it is difficult to
propriate for small a n d / o r on-line computer ap- develop meaningful formulations of the optimiza-
plications. tion problems inherent in the design process.
The economic efficiency of five order picking To add to the difficulty in solving material
systems was investigated by Shimizu I54]. He handling problems, most of the discrete optimiza-
determined the cost per pallet and the cost per tion problems that appear to be worthy of atten-
pick for the five alternatives an6 discussed ap- tion are variations of NP-complete problems. The
propriate applications for each of the alternatives. stochastic aspects of the problem also contribute
Bartholdi and Platzman [4] have studied a num- to the difficulties in obtaining useful solutions to
ber of order picking problems involving both lin- even the simplest design problems.
ear storage and circular storage. The order picking However, it is encouraging to note that people
problem with linear storage involves the in industry are developing feasible solutions to
determination of the optimum sequence of "picks" material handling problems, despite the mathe-
or "visits" on the line; it has a direct computer matical complexities cited above. Furthermore,
analogy involving the sequencing of jobs stored on empirical evidence suggests they are developing
magnetic tape. The order picking problem with reasonably good, if not optimum, solutions. The
circular storage involves the determination of the rules-of-thumb or heuristics used by material
optimum sequence of "picks" on a carousel con- handling engineers represent a "satisficing" ap-
veyor which can be rotated clockwise or counter- proach to both design and operating problems.
clockwise; it has an analogy involving the sequenc- Because of the abilities of an experienced indi-
ing of jobs stored on a disk. Bartholdi and Platz- vidual to resolve some of the more difficult combi-
man have developed exact solutions for a single natoric aspects, of the material handling problem,
orderpicker picking items from multiple carousels, the use of interactive optimization methods is re-
as well as bounds for heuristic procedures. ceiving increased attention. Specifically, color-
316 J.O. Matson, J.A. White / Matertal handling

graphics computer terminals are used to display [3l B.G. Barrett, A further digression on the over-automated
outputs from simulation models and optimization warehouse: some evidence, Interfaces 8 (1977) 46-49.
[4l J.J. Barthoidi, Iii and UK. Platzman, unpublished re-
models. Flexible manufacturing systems, order search papers, Production and Distribution Research
picking systems, automated storage and retrieval Center, School of Industrial and Systems Engineenng,
systems, and warehouse layout problems have been Georgia Institute of Technology, Atlanta, GA.
the subject of stuOies involving the use of color- 15] J.J. Bartholdi, II! and A.S. Rosenthal, Periodic scheduling
graphics computers. of a materials handling system, Prvc. 18th IEEE Con-
ference on Dectswn and Control, Fort L~uderdale, FL
Research is needed in developing large-scale (1979) 548-552.
models of manufacturing systems and warehous- [6] Y. Bassan, Y. Roll and M.J. Rosenblatt, Internal layout
ing systems. The work done on flexible manufac- design of a warehouse, AIIE Trans. 12 (1980) 317-322.
turing systems represents a first step in consider- [7] W,E. Biles and E.Y. Yang, A generalized model and
ing handling, storage, contrc!, and processing fea- simulation analysis of a multiple parallel conveyor system
for flexible manufacturing, in: T.I. Oren, C.M. Delfosse
tures of a m~aufacturing system. Previous research
and C.M. Shub, Eds.. Proc. 1981 Winter Stmulmwn Con-
on warehousing systems has failed to address the ference, Atlanta, GA (1981) 391-399.
integration of receiving, in-bound inspection, [8] Y.A. Bozer, J.S. Shieh and J.A. White, Modeling auto-
movement to storage, storage, retrieval or order mated storage/retrieval systems, TIMS/ORSA Joint Na-
picking, ord,:r accumulation, movement to ship- ttonal Meeting, Detroit, MI (1982).
[9] E.S. Buffa, G.C. Armour, and T.E. Vollmann, Allocating
ping, and shipping; storage, retrieval, and order
faciflties with CRAFT, Harvard Business Rev. 42 (1964)
picking problems are well studied, but the other 136-159,
areas have received ~attle attention. [10] J.A. Buzacott and L.E. Haniftn, Models of automatic
Research contributions in other, seemingly un- transfer lines with inventory banks - a review and com-
related, areas appear to be especially applicable to parison, AIIE Trans. 10 (1978) 197-207.
[! 1] J.A. Buzacott and J.G. Shanthikumar, Models for under-
material handling. For example, there exist~.'a
standing flexible manufacturing systems, AIIE Trans. 12
number of opportunities to apply network and (1980) 339-350.
graph theoretic concepts to the formulation of [12] N.A. DeMars. Computer-aided warehouse layout design
warehouse layout, stock location, and order pick- 30th Annual Material Handling Short Course, Georgia
ing problems. Additionally, much of the computer Institute of Technology, Atlanta, GA (1980).
[13] E.A. Elsayed, AIgortthms for optimal material handling i.n
science research focusing on storage and process-
automatic warehousing systems, Internal. J. Productwn
ing systems has direct application in warehousing Res. 19 0981) 525-535.
and order picking. Research performed by techni- [ 14] R.L. Francis and J.A. White, Fac:hty Layout and Locatwn.
cal geographers and regional scientists concerning An Analytwal Approach (Prentice-Hall, Englewood Cliffs.
spatial relationships may prove to be useful in the N J, 1974).
study of layout problems. [15] S.B. Gersh/vin and O. Berman, Analysis of transfer lines
consisting of two unreliable machines with random
Because of the breadth of the subject of material processing times and finite storage buffers AIIE Trans. 13
handling, as well as the opportunities to apply 0981) 2-11.
results from research performed in other areas, it [16] M.P. Goetschalckx and H.D. Ratliff, Order picking prob-
is not likely that a unified body of knowledge will lems, EURO V/TIMS XXV Conference, Lausanne,
develop that uniquely defines material handling Switzerland (1982).
[17] S.C. Graves, W.H. Hausman, and L.B. Schwarz, Storage-
research. Rather, it is believed that material han- retrieval interleaving in automatic warehousing systems,
dling will continue to provide a wide ranging Management Sci. 23 (1977) 935-945.
number of opportunities for the application of [18] S.C. Graves and B.W. Lamar, A mathematical program-
operational research results. wing procedure for manufacturing system design and
evaluation, Proc. IEEE Conference on Circuits and Com-
puters, Port Chester, NY (1980) 1146-1149.
[19] S.C. Graves and D.E. Whitney, A mathematical program-
References ming procedure for equipment selection and system
evaluation in programmable assembly, Proc. 18th IEEE
[1] G.C. Armour and E.S. Buffa, A heuristic algorithm and Conference on Dectswn and Control, Fort Lauderdale, FL
simulation approach to the relative location of facilities, (1979) 531-536.
Management Scl. 9 0963) 294-309. [20] G. Gregory and C.D. Litton, A conveyor model with
[2] R.D. Armstrong, W.D. Cook and A.L. Snipe, Optimal exponential service times, lnternat. J. Producuon Res. 13
batching in a semi-automated order picking system, J. (1975) !-7.
Operational Res. Sac 30 (1979) 711-720. [21] W.H. Hausman, L.B. Schwarz and S.C. Graves, Optimal
 J.O, Matson, J.A. White / Material handhng 317

storage assignment in automatic warehousing systems, [43] E.J. Muth and J.A. White, Conveyor theory, a survey.
Management Sc:. 22 (1976) 629-638. AIlE Trans. i l (1979) 270-277.
[22] J.L. Heskett, Cube-per-order index - a key to warehouse [44] S.Y. Nor, J.L. Knight, Jr. and G. Salvendy, Effective
stock location, Transportation and Distribunon Manage- utilization of industrial robots - a job and skills analysis
ment 3 0963) 27-31. approach, A I l E Trans. 12 (1980) 215-225.
[23] T.J. Hodgson and T.J. Lowe, Production lot sizing with [45] H. Ohta, GERT analysis of a stogie conveyor s~stem,
material handling cost considerations, A H E Trans. 14 Internat. J Production Res 17 (1979) 405-410.
(1982) 44-51. [46] D.T. Phillips, Simulation of material handling systems
[24] J.J. Hughes, Application of simulation in the design and when and which methodology?, lndusmal Engrg 12 (1980)
development of a flexible manufacturing system, Proc. 64-77.
17th IEEE Conference on Deoslon and Control, San Diego, [47] H.D. Ratliff and A.S. Rosenthal, Order-picking m a rect-
CA (1979). angular warehouse: a solvable case of the traveling sales-
[251 G.K. Hutchinson and B.E. Wynne, A flexible manufactur- man problem, Report No. 81-10, Production and Distri-
ing system, Industrial Engrg. 5 (1973) ! 0-17. bution Research Center, Georgia Institute of Technology,
[26] E. Ignall and A. Silver, The output of a two-stage system School of Industrial and Sytems Engineering, Atlanta, GA
with unreliable machines and limited storage, AIIE Trans. (1981).
9 (1977) 183-188. [48] H.D. Ratliff and T. Tran, Optimum four module pallet
[27] P.S. Jones, A Leust-Cost Equ:pment Selection Techmque for packing, Report No. 82-03, Production and l)lstnbutlon
Dastrtbut:on Warehouses (Stanford Research Institute, Research Center, Georgia Instltme of Technology, School
Menlo Park, CA, 1971). of Industrial and Systems Engineermg, Atlanta, GA ( 19821.
[28] C. Kallina and J. Lynn. Application of the cube-per-order [49] L.W. Rearick, F. Gudan and J F Novotnak, Modeling
index rule for stock location in a distribution warehouse, techniques for manufacturing conveyor systems using
Interfaces 7 (1976) 37-46. GPSS, in: T.I. Oren, C.M. Delfosse and C.M. Shub, Eds.,
[29] Y. Karasawa, H. Nakayama and S. Dohi, Trade-off analy- Proc. 1981 Winter Samulatton Conference, Atlanta. GA
sis for optimal design of automated warehouses, lnternat (1981) 497-504.
,L Systems So. 11 (1980) 567-576. [50] S.D. Roberts and R. Reed. Jr., Optimal warehouse bay
[30] D.A. Kind, Measuring warehouse space utdization, Trans- configurations, A l i e Trans. 4 (1972) 178- 185
portation and Distr:button Management 5 (1965) 22-33. [51] G.M. Sand, Stacker crane product handling system,
[31] D.A. Kind, Elements of space utilization, Transportatzon TIMS/ORSA Joint National Meeting, Miami, FL (1976)
and Distribution Management 15 (1975) 29-34. [52] J.L Santana and L K. Platzman. Real-time scheduling of
[321 E.D. Kooy, Making better use of avadable warehouse an "open warehouse" sorter-palletlzer system, Proc 18th
space, Industr:ai Engrg. 13 (1981) 26-30. IEEE Conference on Decision and Control, Fort Lauder-
[33] E.D. Kooy and D.L. Peterson, Use of the computer m dale, FL (1979) 537-539.
warehouse layout and space planning, Proc. 1971 AIIE [53] L B. Schwarz, S.C. Graves and W H. Hausman, Schedul-
Conference, Boston, MA ( 1971) 13 I - ! 38. ing policies for automatic warehousing systems. ,,lmula-
[34] W.H. Marsh, Elements of block storage design, lmernat. J. tion results, AIIE Trans. 10 (1978) 260-270
Product:on Res. 17 (1979) 377-394. [54] A. Slumlzu, Engineering analysis of automated warehouse,
[35] J.O. Matson and J.A. White, Storage system opumizat~on, Proc 4th International Confereme on Automat:on tn
Report No. 81-09, Production and Distribution Research Warehousmg, Tokyo ( ! 981 ) S 1-5: I - 11.
Center, Georgia Institute of Technology, School of In- [55] A. Smith and P. DeCam, An algorithm to optimize the
dustrial and Systems Engineering, Atlanta, GA (1981) layout of boxes in pallets, J. Operational Res Soc 31
[36] J.O. Matson, J.S. Shieh and J.A. White, Analysts and (1980) 573-578.
design of storage systems, EURO V/TlMS XXV Con- [56] H.J. Steudel, The right pallet cuts costs, lndusmal Engrg 9
ference, Lausanne, Switzerland (I 982). (1977) 14-18
[371 W.L. Maxwell, Solving material handling problems w~th [57] H.J. Steudel, Generating pallet loading patterns a specml
OR, Industrial Engrg. 13 (1981) 58-69. case of the two-dimensional cutting stock problem,
[38] W.L Maxwell and R.C. Wilson, Dynamic network flow Management Sc:. 25 (1979) 997-1004.
modefing of fixed path material handling systems, A l l E [58] D. Stoki~: and M. Vukobratovt~, Simulation and control
Trans. 13 (1981) 12-21. synthesis of manipulator m assembhng techmcal parts. J.
[39] D.J. Medeiros, R.P. Sadowski, D.W. Starks and B.S. Snuth, Dynamw Systems, Measurement, and Control Trans. ASME
A modular approach to simulation of robotic systems, in: 101 (1979) 332-338.
T.l. Oren, C.M. Shub and P.F. Roth, Eds., F'~roc. 1980 [59] R.W. Swain and J.J. Marsh, III, A simulation analysis of
Winter Simulauon Conference, Orlando, FL (1980) 207- an automated hospital materials handling system, A IIE
214. Trans. 10(1978) 10-18.
[40] J.J. Moder and H.M. Thornton, Quantitative analysis of [60] J.A. Tompkins and R. Reed. Jr., An applied model for the
the factors affecting floor space utilization of palletized facdities design problem, lmernat J. Production Re~ 14
storage, J. lndustrial Engrg. 16 (1965)8-18. (1976) 583-595.
[41] E.J. Muth, Modeling and system analysis of multistatlon [61] A.F. Vereschagin and V.L. Generozov, Design of trajecto-
closed-loop conveyors, lnternat. J. Productaon Res. 13 ries of operating device of automatic mampulator, Engrg
(1975) 559-566. Cybernencs 16 (1978) 55-64.
[42] E.J. Muth, A model of a closed-loop conveyor with ran- [62] D.B. Webster and R. Reed, Jr., A material handhng
dom material flow, A H E Trans. 9 (1977) 345-351. system selection model, A i l E Trans. 3 (1971) 13-21.
318 .I.0. Matson, J.A. White / Material handhn$

[63] J.A. White and Y. A. Bozer, Optimum designs of auto- location of stock in a warehouse, AIIE Trans. 9 (1977)
mated storage/retrieval systems, TIMS/ORSA Joint Na- 230-237.
tional Meeting, Washington, DC (1980). [66] H.A. Zollinger, Planning, evaluating, and estimating stor-
[64] J.A. White, N.A: DeMars and J.O. Matson, (~ptimizing age systems, Proc. 1975 Advanced Material Handhng Tech-
storage system selection, Proc. 4th International Conference nology Seminar, Purdue University, West Lafayette, IN
on Automation in Warehousing, Tokyo (1981)$4-5: 1-17. (1975).
[65] H.G. Wilson, Order quantity, product popularity, and the