Content uploaded by Andres Leonardo Carrano
Author content
All content in this area was uploaded by Andres Leonardo Carrano on Jul 14, 2016
Content may be subject to copyright.
On the environmental impacts
of pallet management operations
A. Mazeika Bilbao, A.L. Carrano, M. Hewitt and B.K. Thorn
Department of Industrial and Systems Engineering,
Rochester Institute of Technology, Rochester, New York, USA
Abstract
Purpose – This paper seeks to frame and model the environmental issues and impacts associated
with the management of pallets throughout the entire life cycle, from materials to manufacturing, use,
transportation to end-of-life disposal.
Design/methodology/approach – A linear minimum cost multi-commodity network flow
problem is developed to make pallet-related decisions based on both environmental and economic
considerations.
Findings – This paper presents a review of the environmental impacts associated with pallets by life
cycle stage. The types of materials used to fabricate pallets, the methods by which they are treated for
specific applications, and various pallet management models are described with respect to embodied
energies, toxicity and emissions. The need for companies to understand the cost, durability, and
environmental impact tradeoffs presented by pallet choices is highlighted. The paper introduces a
model to assist in choosing both how pallets are managed and the material they are constructed of that
balances these tradeoffs.
Originality/value – There is limited research on the environmental impact of different management
approaches of large-scale pallet operations. The proposed model and approach will provide companies
seeking to engage in more sustainable practices in their supply chains and distribution with insights
and a decision-making tool not previously available.
Keywords Sustainable pallet operations, Environmental lifecycle design,
Linear cost network flow model, Pallets, Environmental management
Paper type General review
1. Introduction
Supply chains are growing more and more complex. This is due to many factors,
including the expansion of global markets and product storage keeping units, an
increased variety of shipping and distribution modes, and rising expectations from
customers, particularly with respect to service levels and delivery times. At the same
time, companies are striving to make their supply chains more efficient and more
sustainable. One way to do so is to evaluate their shipping and distribution operations.
Pallets, being the most common unit load platform for handling and storing goods, are a
critical component of these operations. Because many pallets are used when producing
and distributing large quantities, the environmental impact associated with the use of a
single pallet is greatly magnified.
The Department of Transportation (BTS, 2009) estimates that transportation
represents roughly 10 percent of the US gross domestic product, or approximately
$1.4 trillion. In 2006, some 8.8 million trucks traveled approximately 263 billion miles.
Freight, in its many forms, accounts for 470 million metric tons of carbon dioxide
equivalent annually (7.8 percent of total US CO
2
emissions), and it contributes about
50 percent of NOx emissions and 40 percent of particulate matter emissions from
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/2040-8269.htm
MRR
34,11
1222
Management Research Review
Vol. 34 No. 11, 2011
pp. 1222-1236
qEmerald Group Publishing Limited
2040-8269
DOI 10.1108/01409171111178765
transportation sources (FHA, 2010; Environmental Protection Agency (EPA, 2006)).
Truck freight accounts for 70 percent of all these emissions. Pallets, the most common
unit load platform, allow the transportation of goods in an efficient, reliable and
seamless way. It is estimated that 80 percent of US trade is carried on pallets (Raballan
and Aldaz-Carroll, 2005). Every year, 500 million new pallets are manufactured and
become part of the large pool (roughly 2 billion) of pallets that are in circulation in the
USA. In the EU some 280 million pallets are in circulation every year. Many of these
pallets are used only a few times and end up meeting a variety of end-of-life scenarios
(e.g. landfill, municipal incineration or downcycling) while others are repaired and
reused many times. As companies set goals to become more sustainable, a thorough
understanding of the environmental impacts of their operations becomes critical.
The ability to control the end of life of the pallets and the associated environmental
impacts of each scenario allows pallet pooling service companies to provide logistics
arrangements that are attractive to those companies seeking to manage their carbon
footprint. However, the complexities of today’s supply chains and the breadth of
environmental impacts pose an interesting challenge to those seeking to engage in
sustainable practices. The challenges will lie on selecting the appropriate pallet type
(i.e. material, durability, etc.) and management structure (e.g. cost, lease vs buy, etc.)
while keeping other aspects in consideration (e.g. toxicity, etc.).
This paper addresses two attributes of a pallet that determine much of its cost and
environmental impact:
(1) how it is managed, which we discuss in Section 2; and
(2) what it is made of, which we discuss in Section 3.
We then propose in Section 4 a method for choosing these attributes in a way that
balances the tradeoffs between cost and environmental impact.
2. Overview of pallet management: open versus closed loops
Pallets are often the unit load of choice when shipping products from the consumer
product manufacturer to the product distributors and/or retailers across the supply
chain. Pallets often represent a significant investment and, once at their destination, an
end-of-life disposition (reuse, recycle, downcycle, incineration or discard) must be
made. The investment in pallets can easily run into millions of dollars for some of the
larger corporations. Thus, the choice of pallet type and management approach has a
significant impact on the bottom line.
In many instances, the choice of pallets and management approach depends
on the type of product and the configuration of the supply chain. For example,
a direct-to-customer retailer of whiteline appliances may choose an open-loop approach
where the ownership of the pallet is transferred to the end-user with the arrival of the
product. In this model, the pallet is not expected to come back to the distributor or
manufacturer and it would likely be disposed to a landfill or given a brief use before
being discarded. These pallets are often referred to as white wood or limited-use pallets
and are chosen by shippers mainly due to the low purchasing cost. This single-use,
open-loop practice, though convenient in some instances, is not sustainable in the long
run and results in tremendous waste and resource consumption. The materials of choice
for these disposable pallets include inexpensive wood (typically a softwood), engineered
Pallet
management
operations
1223
wood (oriented strand board or plywood), structured paper (corrugated fiberboard) or a
mixed product (lumber-reinforced cardboard).
An alternate approach involves the use and implementation of closed pallet loops
(sometimes referred to as pallet pooling). Under this scenario, two options exist:
(1) the seller can contract the use of pallets from a pallet pooling provider
(essentially renting them) or; and
(2) the seller can own pallets but rely on a third party provider to manage them.
Generally, in a closed-loop approach, pallets are collected at a downstream location in the
supply chain (e.g. at the regional distribution center (DC) or retailer) while other pallets
are re-injected at some point upstream in the supply chain (likely at the manufacturing
echelon or one of the consolidation DCs). In this scenario, the reverse logistics associated
with this operation (i.e. backhaul of pallets and pre-position of pallet inventory at the
various echelons) as well as the refurbishing activities gain significant importance.
Pallet pooling done by third party providers involve return centers that collect empties
from many operations and perform inspection, cleaning, repair, sortation and
backhauling. The materials of choice for pallets in closed loops include higher grade
lumber (typically hardwood), reinforced plastic (high density polyethylene HDPE or
polyethylene terephthalate PET), or metals. Figure 1 shows an example of a physical
flow network for pallets in a pooling scenario. A closed loop of pallets allows for a more
efficient recovery and reuse of pallets as well as better management at the end of life.
The services provided by pallet pooling companies usually fall in one of the two
categories: buy/sell or leasing programs. The buy/sell programs use mainly reusable,
higher grade, wood stringer pallets, which are sold to customers, transported with product
through the supply chain (perhaps several times) before being purchased by a local pallet
management facility to be repairedand reused, or recycled (Bejune et al., 2002; IFCO, 2009).
In a leasing program, companies contract the use of a predetermined quantity of pallets
without acquiring ownership. This service usually comes with a structured cost that
comprises a fee associated with the residence time of the pallet in the system, and an issue
fee associated with the costs of repair, cycle count and re-deployment of the pallet.
Figure 1.
An example of a network
of physical flow of pallets
CONSUMER
PRODUCT
MANUFACTURERS
RETAILERS RETURN
CENTERS
PALLET
PROVIDERS
DISTRIBUTION
CENTERS
OTHER
USERS
MRR
34,11
1224
As an example, one of the leading companies in this segment collects, and repairs
approximately 320 million wooden pallets (with over three million of pallet movement
per day) and containers from service centers placed in 75 countries (Raballan and
Aldaz-Carroll, 2005; CHEP, 2008) and for sectors including consumer goods, meat, home
improvement, beverage, raw materials, petro-chemical and automotive industries.
Another leading provider operates a pool of more than 96 million reusable plastic
containers globally, which are used primarily to transport fresh products from
producers to leading grocery retailers.
Finally, a variation of a closed-loop model that is available in certain regions is a
pallet exchange system. The most successful example of such system is the one
established by the Canadian Pallet Council (CPC) in 1968 that currently boasts more than
1,400 members and several pallet services (e.g. short- and long-term rental, one-way
and retrieval programs) (CPC, 2010). In this exchange system, members carry the
responsibility to repair the damaged pallets at certain approved locations.
3. Pallet materials: embodied energy, durability, weight and cost
Approximately, 90 percent of the pallets that are manufactured are made out
of solid wood, with the oak species group being the predominant one within
hardwoods (22.4 percent) and the Southern Pine species group within the softwoods
(7.1 percent) (Bush and Araman, 2008). The other 10 percent include many types of
materials: steel, aluminum, PP, HDPE, PVC, PC, oriented strand board, plywood,
strawboard, corrugated paper, and composites, among others. They all present different
characteristics with respect to cost, durability, weight, sanitization and
decontamination, load rating, stackability, and tolerance for abuse.
The tradeoff between cost and durability of pallets greatly influences the choice of
pallet material. In one report by the World Bank (Raballan and Aldaz-Carroll, 2005),
the costs per trip (one pickup/drop-off cycle with significant travel in between) were
summarized for a few materials: hardwood, softwood and plastic. Although their
analysis is based on assumptions that can be debated, the deductions are nonetheless the
same: softwood pallets can be inexpensive but do not last very long ($6 for a new pallet,
two trips, according to the study) and may not be worth repairing; plastic pallets can last
a long time but their cost is significantly higher (100 trips, $60, according to the study);
while the cost and durability of hardwood pallets tend to fall in between that of softwood
and plastic (but they are typically repaired).
Beyond the tradeoff between cost and durability, there are other issues that can
drive the choice of pallet and the selection of a pallet management strategy.
As organizations work to “green” their supply chains, they will need to take into
account environmentally oriented criteria such as:
.the embodied energy of the materials in the pallet;
.the energy of the pallet manufacturing process;
.the differential emissions that arise during transportation due to the weight
differences between pallets of dissimilar materials;
.potential impacts that arise due to sanitation and sterilization requirements for
pallets; and
.impacts associated with the various end-of-life alternatives available for different
types of pallets (mulching, incineration, landfilling, etc.).
Pallet
management
operations
1225
3.1 Embodied energy
The embodied energy of a material is the energy required to produce a unit of that
material from its raw material ores and feedstocks. Embodied energy is usually
described in terms of energy content per unit weight (e.g. kcal/lb or MJ/kg). This metric
is useful in distinguishing materials that can be synthesized without the investment of
large amounts of energy (e.g. the embodied energy of cast iron ranges from 1.78 to
1.97 kcal/lb) from those that are very energy intensive (e.g. the embodied energy of
platinum is about 12,400,000 kcal/lb) (Cambridge Engineering Selector (CES, 2010)).
The embodied energies for the raw materials generally used to make pallets vary.
High density polyethylene has an embodied energy of 8,320-9,200 kcal/lb, while the
embodied energy of oak (a hardwood used to make durable wood pallets) ranges from
780 to 2862 kcal/lb. The subsequent processing of the raw materials to fashion them
into pallets also consumes energy and therefore adds to the embodied energy of the
finished pallets. Processing HDPE pellets into pallets will require an energy intensive
polymer injection molding process (665-735 kcal/lb) or other thermoforming operation
(e.g. polymer extrusion 262-289 kcal/lb), while transforming oak boards merely requires
simple cutting and assembly which can be done without the investment of much new
energy (51.5-56.9 kcal/lb). The material recycling energy is approximately
2,880-3,190 kcal/lb for HDPE. Both wood and plastic can be combusted for energy
recovery with the net heat of combustion being 4,760-5,010 kcal/lb for HDPE and
2,140-2,310 kcal/lb for oak (CES, 2010).
A full picture of the environmental consequences associated with pallet choice will
need to include the tradeoffs between using more energy intensive plastic pallets over
longer periods of time versus using more, less energy intensive wood pallets for shorter
periods of time.
3.2 Emissions in the use phase (transportation)
Pallets are indirectly responsible for a share of the emissions that are generated as the
pallets and their cargo move through the supply chain. Primary freight transportation
methods (ship, rail, air, and truck) are all fossil fuel based, and heavier pallets will
require more fuel to transport them than lighter pallets. The combustion of the
additional fuel will result in greater emissions of CO
2
, SOx, NOx, and various forms of
particulate matter. Tradeoffs can arise where lighter but less durable pallets could be
preferred to heavier ones because they are responsible for fewer emissions as they
move through a supply chain. Methods are needed to help logistics system designers
understand and evaluate these potential tradeoffs.
3.3 Sanitation and sterilization
Different types of pallets receive different treatment as they move through their
respective supply chains. For example, pallets that are used to transport goods
internationally may be required to undergo sterilization procedures to help prevent the
introduction and spread of invasive species. In particular, the International Plant
Protection Convention (IPPC, 2009) has developed an international standard for
phytosanitary measures (ISPM 15) that describes the internationally accepted measures
that may be applied to wood packaging material by all countries to reduce the risk of
introduction and spread of the pests that may be associated with that material. The
standard sanctions two approaches for helping to ensure that internationally shipped
MRR
34,11
1226
wooden pallets are pest free (the standard does not apply to plastic pallets). The
proposed approaches are:
(1) Heat treatment. Pallets may be heat treated until the core temperature of the
pallet reaches a minimum temperature of 568C for at least 30 minutes.
(2) Fumigation. Pallets may be treated with methyl bromide according to a
schedule that achieves a specified minimum concentration-time product over
24 hours at temperatures and final residual concentrations as specified
(Bush and Araman, 2008).
Note that there may be important environmental impacts to consider with either
approach. Certainly, energy will be required to elevate the temperature of pallets to
568C (1338F) for those pallets that undergo heat treatment. ISPM 15 specifies a number
of processes that can be used to perform this treatment:
(1) kiln-drying;
(2) heat-enabled chemical pressure impregnation; and
(3) microwave treatment.
Clearly, each of these strategies will require the use of energy and will therefore
increase the embodied energy of pallets so treated.
Increasing the embodied energy of pallets is not the only environmental impact
that can arise from sanitation measures. ISPM sanctions fumigation with methyl
bromide as a sanitary measure. When used as a fumigant, methyl bromide gas is
injected into a chamber or under a tarp containing the material to be sterilized. About
80-95 percent of the methyl bromide used for a typical treatment eventually enters the
atmosphere (MBAO, 2010; US EPA, 2010). Methyl bromide is known to be an ozone
depleting material with an ozone depleting potential in the range between 0.2 and 0.5.
Furthermore, methyl bromide is a toxic material. According to the US EPA (2010):
[...] exposure to high concentrations of it can result in central nervous system and respiratory
system failure, as well as specific and severe deleterious actions on the lungs, eyes, and skin.
Exposure to high concentrations has resulted in a number of human deaths.
On the other hand, some plastic pallets are treated with flame retardants, especially
deca-bromine, which is a chemical fire retardant commonly added to the petroleum-based
polymer pallets in order to raise ignition temperature, reduce rate of burning and reduce
time to smoke generation to be equivalent or better than standard wooden pallets
(NWPCA, 2009). There havebeen warnings about the dangers of using palletstreated with
deca-bromine in the hydrocooling process for fruits and vegetables and raised concerns
about the potential carcinogenic effect of deca-bromine (Brindley, 2009).
As companies strive to minimize the overall environmental footprint of their
operations, it will be important to include the potential ozone depletion effects and the
human health impacts in any analysis that attempts to address the fuller environmental
impacts of associated with pallet logistics.
3.4 End-of-life alternatives
Eventually a pallet constructed from any material will come to the end of its useful life.
The methods by which pallets are disposed of when they must be retired can result in
very different environmental impacts.
Pallet
management
operations
1227
Typically, wooden pallets are retired by mulching, incineration, and landfilling.
While mulching is a form of recycling, the material that is recovered is only suitable for
a less demanding application than it was originally used for. This is often referred to as
“downcycling”. Downcycling wooden pallets will require the investment of additional
energy to mulch the wood and transport it to the place where it will be used. This again
adds to the energy embodied in a wooden pallet. On the other hand, landfilling pallet
material offers an opportunity to recover energy since the anaerobic decomposition of
wood generates methane gas which can be captured by modern landfill systems.
Incineration of the pallet material can also be used to generate useful energy, however,
incineration of wood pallets will result in emissions of greenhouse gases, NOx, SOx,
and particulate matter. In addition, combusting pallets that have been treated with
methyl bromide will liberate toxic and irritating chemicals (Cheremisinoff, 1999).
Plastic pallets are generally retired by recycling, incineration, or landfilling. The
recycling of HDPE pallets is actually a downcycling process because the polymer
chains are shortened in the process. The reclaimed HDPE can be mixed with virgin
material to fabricate other plastic products, including pallets.
There is no opportunity to recover energy from HDPE pallets that are landfilled as
HDPE does not decompose in landfills. The plastic from pallets can be used to generate
energy through incineration, however additives like brominated flame retardants can
liberate dangerous substances during the incineration process.
As organizations attempt to more completely address the environmental impacts
associated with pallet logistics, it will be important that they understand the tradeoffs
that arise with respect to the various retirement alternatives that are appropriate for
pallets constructed of differing materials.
4. A proposed model
There are multiple choices for pallet material (hardwood, softwood, or plastic),
management (buy/sell, lease), and end-of-life disposition (reuse, recycle, downcycle,
incineration or landfill), each of which can affect the cost and environmental impact of
possessing and using a pallet. Several studies have addressed different pallet end-of-life
scenarios (Gasol et al., 2008; Corbiere-Nicollier et al., 2001; Bejune et al., 2002;
Buehlmann et al., 2009; Bush and Araman, 2008) which have provided insights on the
specifics of a given end-of-life scenario.
If a decision maker is concerned with only one of the mentioned objectives (e.g. cost
of environmental impact), and wishes to maintain a constant inventory of pallets in the
supply chain, then he/she can use traditional economic analysis methods to make these
choices. For example, suppose cost is the only concern. Each set of pallet choices (e.g. a
leasing a hardwood pallet that will ultimately be discarded, buying a plastic pallet that
will be reused) will have both a lifespan and a cost (in today’s dollars) that reflects both
its initial acquisition and its use over its lifespan. Continuing the example, suppose a
decision maker has determined that 1,000 pallets are needed for 500 consecutive trips
and can choose from the three options presented in Table I. Note that if we assume the
sourcing organization for each pallet option can supply as many as the decision maker
needs, then we can focus on the need for a single pallet. Thus, assuming a trip takes one
week, and dollars are discounted at a rate of 0.05 percent per week, the last row of
Table I can be calculated using present worth analysis, yielding the conclusion that
leasing hardwood pallets is the most economical choice. If environmental impacts are
MRR
34,11
1228
the only concern, then a similar table can be developed, although some thought must be
given to the discount factor used for future environmental impacts.
If a constant inventory of pallets is not desired, a similar, but slightly more
complicated analysis can be done using a tree-based decision analysis framework.
We are interested in how to make these pallet choices when both objectives (cost and
environmental impact) are considered. One approach is to develop a methodology for
converting environmental impacts (such as carbon emissions) to dollars, and then focus
on a single objective, total cost, that includes the dollar cost of those environmental
impacts. An advantage of this approach is that once the methodology for costing
environmental impacts is developed, well-known economic analysis methods can be
used for making decisions. A disadvantage is that developing such methodology adds a
layer of complexity to the decision-making process and will likely require assumptions
that the final decisions prescribed by the analysis may be sensitive to.
Another option is to deal with the two objectives directly and in their own units of
measure. This is the approach we propose, and, to do so, we model these potential pallet
choices with a linear program (a minimum cost multi-commodity network flow
problem). First, we recognize that by choosing a pallet provider, the decision maker
likely also chooses the end-of-life disposition of pallets returned to that provider. Also,
for pallets that are owned by the decision maker’s organization, policies regarding the
end-of-life disposition of pallets are likely already in place and cannot be changed on a
pallet-by-pallet basis. Thus, we limit our model to choosing the material each pallet is
made of and how it is managed, while still considering the environmental impacts
of the end-of-life disposition dictated by that management program. We refer to the
combination of material and management program as a pallet type and denote the set of
possible pallet types as P.
For each pallet type p[P, we model its manufacture, procurement, use, and
disposition with a network similar to what is shown in Figure 3, which adds the
beginning and end-of-life stages for a pallet to Figure 1. For example, for a hardwood
pallet (say p¼1, the arc into “raw material” models the extraction of wood (and other
materials) used to manufacture new hardwood pallets, and we associate with this arc
two attributes ðc1
RM ;e1
RM Þ, the cost and environmental impact of doing so. An example of
calculating e1
RM is given in Table II, where CO
2
footprint is the environmental impact of
concern. There, given the CO
2
footprint per pound of the primary production of the
materials used for a hardwood pallet and their weight (in pounds), e1
RM ¼32:6190. The
arc from “raw material” to “manufactured pallet” models the actual fabrication of a
hardwood pallet. In general, each arc (i, j) in this network, models the transition of a pallet
either from one stage in its life cycle to another or from one type of use to another, and we
associate two attributes, ðcp
ij;ep
ijÞ, with arc (i, j) to represent the monetary cost and
environmental impact of such a transition. Lastly, we note that for some pallet types,
some arcs in Figure 2 may not exist. For example, a plastic pallet may not be eligible
Lease hardwood (pallet
provider discards them)
Lease softwood (pallet
provider discards them)
Buy plastic (will
recycle them)
Cost over lifespan ($) 17.94 11.99 118.51
Estimated life (trips) 25 2 100
Cost for 500 trips ($) 315.53 2,818.36 511.31
Table I.
Three pallet options
Pallet
management
operations
1229
for downcycling, and thus the arc (RC, DEOL) would not exist in the network we use to
model the use of that type of pallet.
Then, to model seasonality in demand and accommodate a multi-period planning
horizon, we discretize time and map the network shown in Figure 2 to a time-space
network with node set N, and arc set A, for each pallet type. A node Iin Nwill be of the
form (s, w), where srepresents a life cycle stage or use phas e as shown in Figure 2, such as
“raw material,” or, “retailers,” and wrepresents a period. For example, if we discretize a
planning horizon of one year into 52 one-week periods, then Nwill contain 52 nodes of
the form (RM, w), where wranges between 1 and 52. Each of these nodes models the
opportunity to extract raw material in week wfor manufacture. Similarly, the time-space
network will contain 52 nodes of the form (MP, w) (CPM, w), ... (RC, w). An arc in Awill
be of the form (i, j)¼((s, w)(s0,w0)), where w0.¼w. For example, Awill contain arcs
of the form ((RM, w)(MP, w 0)) that represent the extraction of raw materials in period
wfor manufacture in period w0. We depict a portion of an example of such a time-space
network in Figure 3. We note that the relationship between wand w0will depend on both
Material Weight (lbs) lbs CO
2
/lb lbs CO
2
Pallet 1
Hardwood oak: medium density 42 0.45 18.88
Low alloy steel: Zn-Cu alloy, fastener wire 3 4.58 13.74
45 32.62
Pallet 2
Hardwood oak: medium density 62 0.45 27.87
Low alloy steel: Zn-Cu alloy, fastener wire 3 4.58 13.74
65 41.61
Table II.
Carbon dioxide footprint
of extracting materials
Figure 2.
The physical flow
of pallets, including
manufacture and
end-of-life
RAW MATERIAL
(RM)
Reuse
Repair
Recycling
MANUFACTURED
PALLET
(MP)
CONSUMER
PRODUCT
MANUFACTURERS
(CPM)
RETAILERS
(RET)
RETURN
CENTERS
(RC)
DISTRIBUTION
CENTERS
(DC)
INCINERATION
END OF LIFE
(IEOL)
LANDFILL
END OF LIFE
(LEOL)
DOWNCYCLING
END OF LIFE
(DEOL)
USE
Downcycling
Landfill
Incineration
MRR
34,11
1230
the discretization of the time and the particulars of the supply chain under consideration.
For example, for a CPM that is six days away from a DC, for the arc ((CPM, w)(DC, w 0)),
w0¼wþ6 if time is discretized into days, but, if time is discretized into weeks,
w0¼wþ1. We model the need for pallets by assuming we know with certainty how
many pallets must flow on the arcs between nodes (CPM, w) and (DC, w 0) and between
nodes (DC, w) and (RET, w 0)for all time pairs (w, w 0) where w0.¼w. In general, for
each arc (i, j) in the time-space network, we assume we know a lower bound l
ij
on the
number of pallets that must flow on that arc.
Because we assume that the management program dictates the end-of-life
disposition of a pallet, we can account for the cost and environmental impact of that
disposition when the pallet is procured. Thus, we include the cost and environmental
impact of end-of-life disposition in the attributes ðcp
ij;ep
ijÞfor arcs from “manufactured
pallet” to “consumer product manufacturers.” If a management program does not
dictate a single end-of-life disposition for a pallet (i.e. 80 percent are repaired
and 20 percent are downcycled), then we include the weighted average of the costs and
environmental impacts associated with the potential dispositions.
We model the finite lifespan of a pallet indirectly, by assuming that a fraction, r
i
p
,
of pallets of type pleaving node iin the time-space network have in fact failed or
become unusable, and thus must leave the system. Indexing this fraction by pallet type
allows us to model the different lifespans associated with pallets of different material.
For example, the failure fraction for softwood pallets can be set much higher than
the fraction for plastic pallets. Thus, for each pallet type, we include arcs of the form
(i, EOL) in the arc set Ato represent pallets failing at node iand thus leaving the
system and reaching their end of life.
Figure 3.
Time-space network
representing physical flow
of pallets, manufacture
and end of life
tt+5t + 2 t + 4t + 3t + 1
MP
RM
DC
CPM
RET
DEOL
RC
LEOL
IEOL
Pallet
management
operations
1231
We model pallet type and quantity decisions with the continuous and non-negative
variables x
ij
p
that represent the number of pallets of type pthat travel on arc (i, j) in the
time-space network. We define Sas the feasible region of the following set of
constraints:
j:ði;jÞ[A
Xxp
ij 2
j:ðj;iÞ[A
Xxp
ji ¼bp
i;i[N;;p[P;ð1Þ
p[P
Xxp
ij $lij ;ði;jÞ[A;ð2Þ
xp
iEOL $rp
i
j:ði;jÞ[A
Xxp
ij ;i[N;;p[P:ð3Þ
The first set of constraints ensures flow balance of pallets within the time-space
network. For nodes that appear early in the planning horizon, the quantities b
i
p
, can
represent the number of pallets of type pthat are already in inventory. The constraints
presented ensure that at the end of the planning horizon a certain number of pallets of
each type are in inventory, which may be too restrictive. To allow more flexibility, one
could, for nodes that appear late in the planning horizon, treat the quantities b
i
p
as
decision variables and include them in the objective function. The second set of
constraints model the demand for pallets during the planning horizon by ensuring that
the total number of pallets that flow across each arc in the time-space network is
greater than some pre-determined lower bound. The last set of constraints model the
failure of pallets of type pthat enter node iby ensuring that, at a minimum, a fraction of
the number of pallets of type pthat enter node idepart on the arc that takes them to
their end of life. Note the third set of constraints also models the opportunity to retire
pallets before they fail. We note that our model does not restrict the decision maker to
choosing one type of pallet management program. We can model such a restriction
with the introduction of a binary variable for each type of pallet management program
that represents whether that particular management program is chosen and ensuring
at most one of those variables is set to 1.
We next define the following two objective functions:
f1ðxÞ¼
p[P
X
ði;jÞ[A
Xcp
ij xp
ij ;
which represents the total monetary cost of the pallet decisions, and:
f2ðxÞ¼
p[P
X
ði;jÞ[A
Xep
ij xp
ij ;
which represents the total environmental impact of the pallet decisions.
To make decisions that jointly consider these two objectives, we propose using this
model in the context of the epsilon-constraint method (Deb, 2001). In general,
with objectives f
1
,f
2
,...f
m
, the epsilon-constraint method repeatedly solves
optimization problems of the form (assuming minimization of all objective functions):
minimize f iðxÞsubject to f jðxÞ#[j;j¼1;2;...;m;j–lx[S
MRR
34,11
1232
where l[{1,2,...,m} and Sis the feasible region, or, in our example a set of feasible
pallet type and quantity decisions. We let P
1
(1
2
) represent the problem where cost is the
objective function we minimize while constraining the environmental impact of our
decisions, and z
1
(1
2
) the cost of the optimal solution to this problem. Then, to generate a
set of solutions, each with a potentially different tradeoff between cost and
environmental impact, we execute Algorithm “ParetoPallet”:
Algorithm Pareto Pallet: Finds pareto – frontier of costs and
environmental impacts
Require: scaling factors
a
1
,...
a
n
Let z2
1¼minimize f
1
(x) subject to x[S. {Calculate least cost set of pallet decisions}
Let z2
2¼minimize f
2
(x) subject to x[S. {Calculate least environmentally harmful set
of pallet decisions}
For i¼1tondo
Set e2¼ð1þ
a
iÞ*z2
2
Solve P
1
(
e
2
) for pallet decisions x*and value z
1
(
e
1
).
end for
Thus, when solving P
1
(1
2
) we are solving the linear program:
minimize
p[P
X
ði;jÞ[A
Xcp
ij xp
ij
subject to
j:ði;jÞ[A
Xxp
ij 2
j:ðj;iÞ[A
Xxp
ji ¼bp
i;i[N;;p[P;
p[P
Xxp
ij $lij ;ði;jÞ[A;
xp
iEOL $rp
i
j:ði;jÞ[A
Xxp
ij ;i[N;;p[P;
xp
ij $0;
p[P
X
ði;jÞ[A
Xep
ij xp
ij #12;
which looks for the least cost set of pallet decisions whose environmental impact is not
too much worse than the least harmful set of pallet decisions.
In words, Algorithm ParetoPallet first finds the set of pallet decisions that are the
least costly and then finds the set of pallet decisions that are the least environmentally
harmful. Then, it repeatedly solves linear programs that model a question similar to the
following: “What is the least-cost set of pallet decisions whose environmental impacts
are at most 120 percent of those of the least environmentally harmful set of pallet
decisions?” Thus, with Algorithm ParetoPallet, we can generate pallet decisions that are
near the Pareto-frontier, or, are Pareto-optimal with respect to cost and environmental
impact; any decisions that cost less must have a greater environmental impact,
Pallet
management
operations
1233
or vice-versa. In addition, Algorithm ParetoPallet will generate multiple decisions that
are Pareto-optimal, allowing the decision maker to choose the tradeoff that fits their
company’s goals. By considering a long planning horizon when constructing the
time-space network, Algorithm ParetoPallet will explicitly capture the tradeoffs
between the cost and environmental impact of manufacturing a pallet and its durability.
With Algorithm ParetoPallet we will also be able to perform what if analysis to
understand this tradeoff, asking questions like “How long must a plastic pallet last to
offset the environmental impact of its manufacture?,” or, “For how long must I need the
use of pallets for the higher cost of plastic pallets to make economical sense?” And of
course, the model is not limited to certain types of pallets. As new materials are
considered for manufacturing pallets, or new manufacturing processes are established,
the model may still be used.
5. Conclusions and future work
In this paper, a comprehensive review of the environmental implications of using
pallets for unit load formation and product delivery are outlined. The types of
materials, treatment methods, and management models are explored with respect to
embodied energies, toxicity and emissions. The need for companies to model the cost,
durability, and environmental impact tradeoffs presented by the pallet choices is
highlighted. We present a method for choosing both how pallets are managed and the
material they are constructed of that balances these tradeoffs. The next steps for this
research are to develop a case study for the method based on the supply chain of a
large, local grocery retailer as well as to explore the ramifications of various design
approaches (e.g. materials substitution or pallet light-weighting strategies) on both
environmental and cost dimensions.
References
Bejune, J., Bush, R., Araman, P., Hansen, B. and Cumbo, D. (2002), “Recycling behind most
growth in pallet production”, Pallet Enterprise, 3 September, available at: www.
palletenterprise.com/articledatabase/view.asp?articleID¼768 (accessed July 3, 2009).
Brindley, C. (2009), “How safe is deca bromine? FDA statement raises toxicity concerns for some
plastic pallets”, Pallet Enterprise, July 1, available at: www.palletenterprise.com/
articledatabase/view.asp?articleID¼2886 (accessed December 14, 2010).
BTS (2009), National Transportation Statistics 2009, US Department of Transportation, Bureau
of Transportation Statistics, Washington, DC, available at: www.bts.gov (accessed July 3,
2010).
Buehlmann, U., Bumgardner, M. and Fluharty, T. (2009), “Ban on landfilling of wooden pallets in
North Carolina: an assessment of recycling and industry capacity”, Journal of Cleaner
Production, Vol. 17 No. 2, pp. 271-5.
Bush, R.J. and Araman, P.A. (2008), “Updated pallet and container industry production and
recycling research”, Internal White Paper.
CES (2010), CES Level 3 Eco-selector Materials Database, Version 5.1.0, Cambridge Engineering
Selector, Cambridge.
CHEP (2008), Building a Sustainable Supply Chain, 23rd ed., Commonwealth Handling
Equipment Pooling, Orlando, FL.
Cheremisinoff, N.P. (1999), Handbook of Industrial Toxicology and Hazardous Materials, CRC
Press, New York, NY.
MRR
34,11
1234
Corbiere-Nicollier, T., Gfeller-Laban, B., Lundquist, L., Leterrier, Y., Manson, J.A.E. and Jolliet, O.
(2001), “Life cycle assessment of biofibres replacing glass fibres as reinforcement in
plastics”, Resources Conservation and Recycling, Vol. 33 No. 4, pp. 267-87.
CPC (2010), available at: www.cpcpallet.com (accessed December 5, 2010).
Deb, K. (2001), Multi-objective Optimization Using Evolutionary Algorithms, Wiley, New York, NY.
EPA (2006), US Greenhouse Gas Inventory Report, Annex 2, US Environmental Protection
Agency, Washington, DC.
FHA (2010), Freight Facts and Figures, US Department of Transportation, Federal Highway
Administration, Washington, DC.
Gasol, C.M., Farreny, R., Gabarrell, X. and Rieradevall, J. (2008), “Life cycle assessment
comparison among different reuse intensities for industrial wooden containers”,
International Journal of Life Cycle Assessment, Vol. 13 No. 5, pp. 421-31.
IFCO (2009), “Factors impacting the environmental sustainability of pallet programs”, Internal
White Paper, IFCO Systems, Atlanta, GA.
IPPC (2009), available at: www.ippc.int/file_uploaded/1285321495_ISPM_15_Revised_2009_E.
pdf (accessed December 5, 2010).
MBAO (2010), “The US EPA methyl bromide phase out website”, available at: http://mbao.org
(accessed December 5, 2010).
NWPCA (2009), “Cost and performance tradeoffs of pallets and racks”, Pallet Enterprise,
available at: www.nwpca.com/_INTLRegulations/PlasticPalletsFireHazardNews.pdf
(accessed September 3, 2009), Pallets and Packaging 101.
Raballan, G. and Aldaz-Carroll, E. (2005), “How do different standards increase trade costs?
The case of pallets”, World Bank Policy Research Working Paper No. 3519, The World
Bank, Washington, DC, February.
US EPA (2010), “The phaseout of methyl bromide”, available at: www.epa.gov/ozone/mbr/
(accessed December 5, 2010).
Further reading
Lacefield, S. (2004), “What is more ‘palletable’ – renting or owning?”, Logistics Management,
April.
About the authors
A. Mazeika Bilbao is a Graduate Assistant in the MSc degree in Sustainable Engineering
program at Rochester Institute of Technology (RIT) in New York. She received her BS degree in
Industrial Engineering from Universidad Catolica Andres Bello in Venezuela. Her professional
interests include sustainability, life cycle analysis, renewable energies, supply chain modeling,
and lean manufacturing. She has received the Outstanding Graduate Student Award from the
Industrial and Systems Engineering Department at RIT, and is a member of the Phi Kappa Phi
Honor Society, IIE, SWE, and SHPE.
A.L. Carrano is an Associate Professor of Industrial and Systems Engineering and the
Director of the Toyota Production Systems Lab at the RIT in New York. He holds a PhD and an
MSc from North Carolina State University and a BS from Universidad Catolica Andres Bello in
Venezuela. His teaching interests are in the areas of production systems, material handling and
design. His research interests lie in the areas of sustainable product design, surface metrology
and manufacturing processes.
M. Hewitt, BS 1995 University of Michigan, MSE 1998 University of Michigan and PhD 2009
Georgia Institute of Technology, is an optimization and supply chain specialist. His current
Pallet
management
operations
1235
research focuses on developing hybrid methods that integrate exact optimization and heuristic
search techniques. He is particularly interested in problems in the transportation and supply
chain management domains and has developed innovative techniques for solving large-scale
optimization problems in both freight transportation and maritime inventory routing. His work
has assisted the decision making of companies such as Exxon Mobil, Saia Motor Freight, and
Yellow Roadway. His current work includes improving the scheduling and routing of home
health care nurses and using machine learning techniques to solve discrete optimization
problems. Before entering the PhD program at Georgia Tech, Mike worked as a software
engineer, contributing to the development of software to support consumer set-top boxes and
content delivery to LED signs in mass transit stations. M. Hewitt is the corresponding author and
can be contacted at: mrheie@rit.edu
B.K. Thorn is an Associate Professor in the Industrial and Systems Engineering Department
at the RIT in New York. He received his BS in Industrial Engineering from RIT, and his MS and
PhD from Georgia Tech. His research interests include applied statistical methods, sustainable
product and process design as well as life cycle analysis. He is a member of IIE, ASEE and ESW.
Dr Thorn is co-advisor of the ESW student chapter at RIT and oversees, along with Professor
Andres Carrano, a group of graduate students with common research interests in sustainability.
MRR
34,11
1236
To purchase reprints of this article please e-mail: reprints@emeraldinsight.com
Or visit our web site for further details: www.emeraldinsight.com/reprints
