Professional Documents
Culture Documents
Bertazzi2002 (Multi Products and Multi Customer)
Bertazzi2002 (Multi Products and Multi Customer)
Transportation Science
Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org
This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.
The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.
© 2002 INFORMS
INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
Deterministic Order-Up-To Level Policies
in an Inventory Routing Problem
Luca Bertazzi • Giuseppe Paletta • M. Grazia Speranza
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
multiperiod models with deterministic demand, and determines a lower and an upper level of the inven-
infinite-horizon models with deterministic demand. tory of each product and can be visited several times
An example of single-period models with stochas- during the time horizon. Every time a retailer is vis-
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
tic demand can be found in Federgruen and Zipkin ited the quantity delivered is such that the maximum
(1984). In this model, the quantity of product to ship level of inventory is reached. This inventory policy
to each retailer is determined on the basis of the level is inspired, in a deterministic setting, by the classical
of the inventory at the retailer. Then, the retailers order-up-to level policy, widely studied in inventory
are assigned to the vehicles and the routes are deter- theory (see Axsäter and Rosling 1994 for an overview
mined. The multiperiod models with deterministic of inventory policies in multilevel systems). A unique
demand are deterministic models in which several decision maker, typically, a logistics manager, has to
shipments can be performed over a time horizon. In determine a shipping policy that minimizes the sum
these models, simpler models are often used as sub- of transportation costs and inventory costs both at the
problems. For instance, in Dror and Ball (1987) single- supplier and at the retailers. A shipping policy con-
period models are used as subproblems, while in sists of determining for each delivery time instant the
Bertazzi et al. (1997) direct shipping models are used. set of retailers to visit, the quantity of each product
Finally, in the infinite-horizon models with determin- to ship to each retailer, and the route of the vehicle.
istic demand the product is absorbed by each retailer The constraints of the problem guarantee that in each
route the capacity of each vehicle is not exceeded and
at a given constant rate and the problem is to deter-
that the level of the inventory both at the supplier
mine an infinite-horizon shipping policy that mini-
and at each retailer is never lower than the minimum
mizes the sum of inventory and vehicle-routing costs.
level. This problem, referred to as deterministic order-
Examples of these models can be found in Anily
up-to level inventory-routing problem, is obviously
and Federgruen (1990) and in Chan and Simchi-Levi
NP-hard, because it reduces to the TSP in the class of
(1998).
instances in which the time horizon is one, the inven-
We study a multiperiod model with deterministic
tory costs are zero, the capacity of the vehicle is infi-
demand in which a set of products is shipped from
nite, and all the retailers need to be served.
a common supplier to several retailers. Each prod-
The scope of this paper is twofold. On one hand,
uct is made available at the supplier and absorbed by
we aim at solving the above problem. On the other
the retailers in a deterministic and time-varying way.
hand, we intend to study the impact of the objective
By deterministic way we mean that the quantity of function on the problem solution. Because the prob-
each product that will be made available at the sup- lem is very complex and the exact solution would be
plier and the quantity that will be absorbed by the impractical, we propose a constructive heuristic algo-
retailers are known in each discrete time instant. By rithm in which at each iteration a retailer is inserted
time-varying way we mean that the quantity of each in the solution. For each retailer, the heuristic builds
product made available and absorbed in each time a network to represent the incremental costs due to
instant can be different from the one made available the insertion of the retailer in the solution. The short-
and absorbed in a different time instant. The prod- est path of the network identifies a “min cost” policy
ucts are shipped from the supplier to the retailers by for adding the retailer to the solution. To evaluate
a vehicle of given transportation capacity and cost. the impact of the objective function on the problem
Shipments can be performed only in the discrete time solution, we consider different objective functions that
instants (delivery time instants) that belong to a given include some of the cost components only: the trans-
time horizon. For each product, a starting level of the portation cost only, the inventory costs at the retailers
inventory is given both for the supplier and for each only, and the sum of the transportation cost and the
retailer, and the level of the inventory at the end of the inventory cost at the supplier. In this way we obtain
time horizon can be different from the starting one. three variants of the problem, never studied in the lit-
Therefore, the problem is not periodic. Each retailer erature to the best of our knowledge. We try here to
give a partial answer to the question of how impor- An inventory cost is charged both at the supplier
tant the coordination issue is, evaluating which of the and at the retailers. If we denote by hi the unit inven-
solutions is closer to the solution in which the total tory cost at node i ∈ , then the total inventory cost
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
cost is minimized, the impact of each cost on the total over the time horizon can be computed as follows.
cost, and how the objective function affects each type At the supplier 0 the level of the inventory at time
of cost. To have a better understanding of the model, t + 1 is given by the level at time t, plus the quantity
in this paper we investigate the case of a single prod- of product made available at time t, minus the total
uct and a single vehicle. The model and the proposed quantity shipped to the retailers at time t, that is
heuristic can be extended easily to the case of multi-
Bt+1 = Bt + r0t − xit
ple products and vehicles. i∈
The paper is organized as follows. In §1 the prob-
lem is described and in §2 the heuristic algorithm is where B0 (the starting level of the inventory) is given,
presented. The variants of the problem, with different r00 = 0 and xi0 = 0, ∀ i ∈ . Therefore, the total inven-
objective functions, are introduced in §3. Finally, in §4, tory cost at the supplier is t∈ h0 Bt , where = ∪
the computational results obtained on randomly gen- H + 1. At each retailer i ∈ , the level of the inven-
erated problem instances are shown and discussed. tory at time t + 1 is given by the level at time t plus
the quantity of product shipped from the supplier to
the retailer i at time t minus the quantity of product
absorbed at time t; that is,
1. Problem Description
We consider a logistic network in which a prod- Iit+1 = Iit + xit − rit
uct is shipped from a common supplier 0 to a set
= 1 2 n of retailers over a given time horizon where Ii0 (the starting level of the inventory) is given,
H . In each discrete time instant t ∈ = 1 2 H and ri0 = xi0 = 0. Therefore, the total inventory cost at
a quantity rit of the product is absorbed at the retailer the retailer i is t∈ hi Iit . The time instant H + 1 is
i ∈ and a quantity r0t is made available at the sup- included in the computation of the inventory cost to
plier. Each retailer i ∈ defines a minimum level Li take into account the consequences of the operations
and a maximum level Ui of the inventory of the prod- performed at time H .
uct. If retailer i is visited at time t, then the quantity xit The problem is to determine for each retailer i ∈
of product shipped to retailer i is such that the level a set i of delivery time instants, and for each time
of the inventory in i reaches its maximum level Ui instant t ∈ a route t that visits all the retailers
(order-up-to level policy). More precisely, if we denote served at time t such that the sum of the transporta-
by Iit the level of the inventory at retailer i at time t, tion cost and of the inventory cost both at the supplier
and at the retailers is minimized and the following
then xit is either equal to Ui − Iit if a shipment to i is
constraints are satisfied:
performed at time t, or equal to 0 otherwise.
(1) Capacity Constraints. They guarantee that the
Shipments from the supplier to the retailers can
total quantity of the product loaded on the vehicle is
be performed in each time instant t ∈ by a vehi-
not greater than the transportation capacity:
cle with given capacity C, and routing is allowed. In
each route the vehicle visits the retailers that must be
xit ≤ C t ∈ (1)
served at the same time instant. We assume that any i∈
combination of customers can be visited in a single
(2) Stockout Constraints at the Supplier. They guaran-
delivery time instant. The transportation cost cij from
tee that the level of the inventory Bt is nonnegative in
i to j, with i j ∈ = ∪ 0, is known. Therefore,
each time instant t ∈ :
given a route, the corresponding transportation cost
is simply obtained by summing up the cost of the arcs xit ≤ Bt t ∈ (2)
that belong to the route. i∈
1
H t∈ rit . If there exist retailers with the same
Iit ≥ Li t ∈ i ∈ (3) ratio, sort them in the nonincreasing order of
Ui − Li . Rename the retailers accordingly.
Note that the problem in which several products
have to be shipped from the supplier to the retailers 1. Start.
can be handled easily by this model. It is sufficient to For s = 1 2 n
replace each retailer i with a set of retailers, one for — Determine for the retailer s ∈ a set s of
each product absorbed by the retailer i, setting to zero delivery time instants by using the procedure
the transportation cost between nodes of the same Assign.
set and setting to the transportation cost between the — For each time instant t ∈ s insert the retailer
retailers the transportation cost between nodes of dif- s in the route t traveled by the vehicle at
ferent sets. time t by using the procedure Insert.
2. Improve.
Let TC be the total cost of the solution.
2. A Heuristic Algorithm (a) For s = 1 2 n
We propose a two-step heuristic algorithm to solve
For i = n n − 1 1 and i = s
the order-up-to level inventory-routing problem. In
— t = t , ∀ t ∈ .
the algorithm the retailers are ranked in the nonde-
— Remove the retailer s from the routes
creasing order of the average number of time units
t by using the procedure Remove. Do
needed to consume the quantity Ui − Li and the retail-
the same for the retailer i.
ers with the same number of time units are ranked in
— Determine for i a new set of deliv-
the nonincreasing order of Ui − Li . In the initialization
ery time instants by using the proce-
phase of the algorithm, referred to as Start, a feasible
solution of the problem is built by an iterative pro- dure Assign and insert the retailer in the
cedure that inserts a retailer at each iteration. When routes t by using the procedure Insert.
retailer i is considered, a set of delivery time instants Do the same for the retailer s.
is determined by solving a shortest-path problem on — Let TC be the cost of the obtained solu-
an acyclic network in which every node is a possi- tion. If TC < TC, then adopt the new
ble delivery time instant (procedure Assign). Then, for solution.
each of the selected delivery time instants, the retailer (b) If a new solution has been adopted for at
is inserted in the route traveled by the vehicle by least a pair of retailers, then go to (a).
applying the well-known rule of insertion at cheap-
est cost (procedure Insert). In the second phase of the 2.1. Determining the Delivery Time Instants
algorithm, referred to as Improve, the current solu- In this section we describe the procedure Assign
tion is improved iteratively. At each iteration, a pair which is used to determine a feasible set of delivery
of retailers is temporarily removed from the current time instants for each retailer s. This procedure works
solution by using the procedure Remove. Then, each of on an acyclic network Gs s s s s , in which
the removed retailers is inserted in the current solu- each element of the set s is a node that corresponds
tion as in the procedure Start. Finally, if this reduces to a discrete time instant between 0 and H + 1 and
the total cost, then the solution is modified accord- each element askt of the set s is an arc that exists if no
ingly. The second step of the algorithm is repeated as stockout occurs in s whenever s is not visited between
long as an improvement in the total cost is reached. k and t. Therefore, each path on the network between
The algorithm can be formally described as follows. 0 and H + 1 represents a set of delivery time instants
for s that satisfy the stockout constraints (3). Each ele- to be inserted between two of the nodes of the route
s
ment qkt of the set s is a weight on the arc askt that t ; given that in the algorithm the rule of insertion at
represents the quantity of product to deliver at time t, cheapest cost is used (see, for instance, Rosenkrantz
s
et al. 1977), then s would be inserted between the
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
any retailer, then each route t , t ∈ , is empty and Once the weight pkt s
is computed for each arc
the level of the inventory Bt at the supplier is equal to askt∈ s , the procedure determines the shortest path
the level obtained if no shipments occur up to time t, between 0 and H + 1 by using an algorithm for acyclic
that is, Bt = B0 + tj=1 r0j−1 , for each time instant t ∈ . networks (see, for instance, Hu 1982) to obtain a set
s
Given the partial solution, the weight pkt is computed of delivery time instants for s that allows us to “min-
as the sum of three components. The first one, c̃ts , is imize” the total cost. Finally, the procedure includes
the estimate of the variation in the transportation cost in the set s of the selected delivery time instants for
obtained if the retailer s is served at time t. This esti- s the intermediate nodes that belong to the shortest
mate is 2c0s if no retailers are visited at time t in the path.
partial solution, i.e., if t = . Otherwise, c̃ts is com- The procedure Assign can be formally described as
puted by taking into account that the retailer s has follows.
• If the length of the shortest path is finite, then The insertion of the retailer s in the route t implies
include in the set s the intermediate nodes that an increase in the total quantity of the product loaded
belong to the shortest path; otherwise, the algorithm on the vehicle equal to qkt s
, a variation in the trans-
is not able to find a feasible solution. portation cost equal to either 2c0s if the route t was
Note that the total quantity Qts of product shipped
empty before inserting s or ci∗ s + cs sui∗ − ci∗ sui∗ oth-
to s up to each delivery time instant t, 1 ≤ t ≤ H , is
erwise, and a reduction of the level of the inventory
independent of the path between 0 and t selected on
at the supplier equal to
the network Gs s s s s . In fact, if t1 t2 tn
are the delivery time instants selected up to time t, s
Bj+1 = Bj+1 − qkt j = t H
with t = tn , then the total quantity Qts shipped to s up
to time t is q0ts 1 + qts1 t2 + · · · + qtsn−1 tn , that is equal to Us −
t 1 Let us now describe the procedure Remove used
Is0 + tj=1 rsj−1 , as q0ts 1 = Us − Is0 + j=1 rsj−1 and qtsm tm+1 =
tm+1 s
during the algorithm to remove the retailer s from
j=1+tm rsj−1 , 1 ≤ m < n. Therefore, Qt is independent the route t . Two different situations can happen,
of the selected delivery time instants t1 t2 tn−1 .
depending on the fact that s is the only retailer visited
Moreover, note that the total quantity K s of prod-
in the route or not before removing it. Let prs be the
uct shipped to s during the time horizon depends on
predecessor to the retailer s in the route t .
the last delivery time instant t̂ selected for s; in fact,
K s = Qt̂s . Therefore, K s can be different from the total Procedure Remove
quantity t∈ rst of product absorbed from s during
the time horizon. If t = 0 s 0, then remove the arcs 0 s
and s 0.
2.2. Inserting and Removing a Retailer
Else
In this section we describe the procedures Insert and
—Remove the arcs prs s and s sus.
Remove which are used during the algorithm to insert
—Introduce the arc prs sus.
and to remove, respectively, a retailer s from the route
t traveled by the vehicle at time t. Let us first con-
The decrease in the total quantity of the product
sider the procedure Insert. As described in the previ- s
ous section, two different situations can happen when loaded on the vehicle is qkt , while the variation in the
the retailer s has to be inserted in the route t . The transportation cost is −2c0s if only the retailer s was in
first one happens when the route t is empty; in this the route before to remove it, and is cprs sus −cprs s −
case, the insertion of the retailer gives a route com- cs sus otherwise. Finally, the level of the inventory at
posed of only the arcs 0 s and s 0. The second one the supplier becomes
happens when the route t already contains some
s
retailers; in this case, the rule of insertion at cheap- Bj+1 = Bj+1 + qkt j = t H
est cost described in the previous section is used. The
procedure can be formally described as follows.
Procedure Insert 3. Variants of the Problem
If t = , then t = 0 s 0. In this section we consider variants of the problem
Else described in §1, in which the aim of the decision
—Select the retailer i∗ such that maker is not to minimize the overall cost but to min-
imize some of the cost components only. All the vari-
i∗ = arg mincis + cssui − cisui ants share the same set of data and the same set of
i∈t
constraints of the problem described above. The only include the estimate c̃ts of the variation in the trans-
difference is in the objective function. portation cost only, and the total cost TC is computed
In the first variant the aim of the decision maker is accordingly.
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
to minimize the sum of the inventory cost at the sup- The third variant puts the focus on the retail-
plier and of the transportation cost only, without any ers’ costs and considers in the objective function the
regard to the inventory costs at the retailers. This hap- inventory costs of the retailers only. The optimal solu-
pens when the shipping policy is defined by the sup- tion here tends to serve each retailer when the inven-
plier who organizes the routing and determines at its tory reaches the minimum level, as stated by the clas-
discretion when to visit each retailer, under the con- sical order-up-to level policy. The only reason why
dition that the level of the inventory of each retailer the optimal solution may happen to serve a retailer
is never lower than the given minimum. In this case, before it is necessary is the capacity of the vehicle or
the supplier has a global view of his own internal the stockout constraints at the supplier. Because an
costs but does not care about the retailers’ costs. This optimal solution only determines the delivery time
may be due to a lack of interest in the retailers’ costs instants and leaves the routing undetermined, we
when supplier and retailers are different companies; consider the transportation cost as a second objec-
instead, it may be simply due to a lack of coordina- tive. This variant corresponds to the situation where
tion in the case in which the supplier and the retailers the retailers can determine the delivery times and
are part of the same company. This first variant is the routing is optimized afterwards. This hierarchi-
NP-hard because it has the TSP as a particular prob- cal problem is again NP-hard and can be heuristi-
lem and can be heuristically solved by a modified cally solved by computing in the procedure Assign the
weights pkts
as the estimate Ikts of the variation in the
version of the algorithm described in the previous sec-
inventory cost at the retailer s only and the total cost
tion. The only required change is in the procedure
s TC accordingly.
Assign where the weights pkt have to be computed
The original problem and the above variants model
here as the sum of the estimate c̃ts of the variation in
s of the different organizational situations of a company, with
the transportation cost and of the estimate B kt
a different emphasis on what is important in the
variation in the inventory cost at the supplier only.
objective of the decision. In the following section, we
The total cost TC is computed accordingly.
compare the solutions obtained in the various cases,
A second interesting variant is obtained when the
trying to derive some conclusions about the impact
objective function includes only the transportation
of the objective on the solution, both in terms of total
cost. This may happen when the transportation is out-
cost and in terms of single components of the total
sourced and the agreement between the supplier and
cost.
the transportation company defines the basic princi-
ples of the service, namely that the stockout situa-
tions have to be avoided. However, the transportation 4. Computational Results
company can organize the deliveries on the basis of The heuristic algorithm described in §2 has been
its own costs. Then, the inventory costs at the sup- implemented in Fortran and used to solve the prob-
plier and at the retailers follow from the shipping lems described in §§1 and 3 in a set of computa-
policy of the transportation company. In the case in tional experiments. We first describe how the random
which the whole system is owned by the same com- instances have been generated. Then we show a
pany, the focus on the transportation cost only may series of computational results to analyze the solu-
again be simply due to a lack of global view and tion obtained for the problem in which the total cost
to a lack of coordination. This variant remains NP- is minimized and to compare the solutions obtained
hard since the routing problem remains to be solved by the different variants of the problem. Finally, we
here and can be heuristically solved by modifying the computationally evaluate the quality of the solution
s
weights pkt of the procedure Assign. Here, the weights generated by the heuristic algorithm.
4.1. Generation of the Instances Two hundred and forty instances have been gener-
The instances have been generated on the basis of the ated on the basis of the data presented in §4.1; in
following data: particular, 10 instances have been randomly gener-
Number of retailers n: 50;
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
Figure 1
retailers, and inventory cost at the supplier, respec- due to the fact that the supplier tends to ship more
tively. In each class of instances the inventory cost frequently.
at the retailers is the main part of the total cost, fol- Figure 1 allows us to evaluate the impact of each
lowed by the inventory cost at the supplier and then cost component on the total cost in the classes of
by the transportation cost. In the first two rows of instances described above. The figure is composed
the table the average costs obtained in the classes of of three figures: The first one (1a) compares the classes
instances with low and high transportation cost are of instances with low and high transportation cost, the
compared. The instances with high transportation cost second one (1b) the classes of instances with low and
have slightly lower inventory cost at the retailers but high inventory cost at the retailers, and the third one
higher inventory cost at the supplier. This is an effect (1c) the classes of instances with low and high inven-
that can be explained by observing that an increase of tory cost at the supplier. In each figure the comparison
between the two classes of instances is performed by
the transportation cost tends to reduce the number of
computing, for each class, the percentage of each cost
times a retailer is served and, therefore, to reduce the
component in the total cost. The first figure shows
transportation cost and to increase the inventory cost
that in the class of instances with high transportation
at the supplier. Rows 3 and 4 allow us to compare
cost the impact of the inventory cost at the retailers
the average costs obtained in the classes of instances
on the total cost is significantly lower with respect
with low and high inventory cost at the retailers. The
to the class with low transportation cost; the second
results show that the instances with high inventory
figure shows that in the class of instances with high
cost at the retailers have lower transportation cost, inventory cost at the retailers the impact of the trans-
but higher inventory cost at the supplier. This is due portation cost on the total cost is significantly lower
to the order-up-to level policy; in fact, contrary to with respect to the class with low inventory cost at
the classical inventory models, here the inventory cost the retailers; finally, the third figure shows that the
at the retailers is minimized by serving the retailers impact of the inventory cost at the retailers and of the
rarely over time. Finally, in rows 5 and 6 the aver- transportation cost does not significantly change in
age costs obtained in the classes of instances with low the classes of instances with low and high inventory
and high inventory cost at the supplier are compared. cost at the supplier.
The results show that the instances with high inven- In Table 2 a sensitivity analysis on the total cost in a
tory cost at the supplier have higher transportation given instance is presented. Each row of the table cor-
cost and higher inventory cost at the retailers. This is responds to a parameter of the instance and shows the
Table 2 Sensitivity Analysis of the retailers. This behavior reflects the fact that the dif-
Total Cost ferent cost components have a significantly different
− 04 + 04 importance in the total cost. The last two rows of the
table show the variations obtained in the total cost by
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
Table 4 Percent Increase of the Costs the transportation cost on one hand and the inven-
IS + IR + T IR T IS + T
tory cost at the supplier on the other hand. When the
inventory cost at the supplier is not included in the
Total cost 01 140 86 84 objective function (see problems IR and T ) the retail-
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
Figure 2
Figure 3
Table 6 Sensitivity Analysis of the Number of Visits each discrete time instant all the retailers on the basis
IS + IR + T IS + T
of an optimal route. The second policy, referred to as
Latest, is to visit in each delivery time instant t the
− 04 + 04 − 04 + 04
Downloaded from informs.org by [130.75.172.160] on 11 March 2015, at 09:47 . For personal use only, all rights reserved.
Received: December 2000; revision received: September 2001; accepted: September 2001.