Sustainable Development of The Supply of Grain

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 14

Sustainable Development of the Supply of Grain

J.L. Corner, LR. Foulds

Department of Management Systems, University of Waikato, Hamilton, New Zealand

Abstract - We develop a silo location for the sustainable

development of the supply of grain. The model is designed to

be utilized in the storage and transport of grain in relatively

large, rural grain producing areas. All current modeis

known to the authors do not address issues related to the

sustainability of such supply chains. The main purpose of the

present paper is to present an innovative medel in order to

provide new insights into important aspects of the

sustainability of such supply chains. The model incorporates

various transportation and silo construction capital

investment costs. Similarities are made between existing

modeis and that developed. The paper contributes to

industrial engineering management by providing an

interesting new location model. This provides new insights

into important aspects of practical grain supply chain

development.

Keywords: Models, networks, supply chain management,

sustainability

1. INTRODUCTION

We examine the question of where to locate silo

complexes for the storage of grain in relatively large rural,

geographical grain producing areas, such as in parts of

Australia, the United States, and the Russian Federation.


Às is common in these and similar regions, the grain is

transported from farms, stored temporarily in the silos,

and then transported again to a number of given

destinations such as ports, processing plants, and markets.

M, as is often the case, production and demand patterns

change significantly within a given planning horizon, it is

far from clear that the silos will all continue to be located

at sites that enable efficient overall grain transportation.

However, as far as the authors are aware, the issue of the

long term sustainability of grain supply is not addressed

by any of the existing location models. The purpose the

present paper is to attempt to fill this gap.

We adopt a “green fields” approach to the question of

silo location with a view to long term sustainability of

grain supply. À dynamic network flow location model is

presented that incorporates transportation, capital

investment and construction costs. An expesition of the

literature on dynamic network models is provided and it

shown how the present paper fits in with previous

knowledge. The model provides new insights into

important aspects of certain practical location problems.

We end the paper with our conclusions and a summary,

We begin by surveying existing general network models

and solution techniques for them.

0-7803-8519-5/04/820.00 (€) 2004 IEEE

0. A SURVEY OF GENERAL NETWORK

MODELS
Classical Network Optimization Models

The carliest network optimization models and

solution techniques concern what is known as the

transportation problem (TP) ([1], [2], [3]). The models to

be developed later in this paper are based on a variation of

this problem, known as the transshipment problem (TSSP)

([4], [5]). These and other network problems are special

cases of a generic network model, termed the minimum

cost flow problem (MCFP). ([6], [5], [7], [4], and [8]). As

well as the TP previously mentioned, other special cases

of the MCFP include the shortest path problem (SPP) ([9]

and [10]), the maximum flow problem (MFP) ([7] and

[11]), and the assignment problem (AP) [12].

Dynamic Networks

Al the network flow models discussed above are

static, in the sense that none of them take temporal

considerations into account. Às is well known, these

models are useful in modelling many types of human

endeavour. However it is sometimes essential to pay

proper regard to the time element, such as in scheduling,

production, capacity, and inventory management,

economics, energy, transport, logistics, and traffic

engineering. Many scenarios that change over the time

periods of a given planning horizon, including those just

given, can often be usefully modelled as dynamic (rather

than static) flow networks.

Dynamic flow networks can be thought of as


duplicate copies of an underlying static network, where

each copy corresponds to a distinct time period of the

planning horizon. Linkages (in the form of arcs) between

the static network copies are used to model, via a master

network, the changes that occur over time in the system

being modelled. As all the nodes of each static sub-

network are associated with a particular instant in time,

the master network is termed a time-expanded flow

network, or a dynamic flow network. We use the latter

term throughout the rest of this paper.

Dynamic network flow models and solution

techniques have been in use for at least 50 years. Early

examples include tanker scheduling [13], maximum flows

([14], [15]), airline scheduling [16], building evacuation

[17], and vehicle fleet sizing [18]. A general introduction

to the field has been given by [19] and a comprehensive

exposition of the theory and applications of dynamic

networks to has been provided by [20]. We now go on to

1226
develop a comprehensive model of grain silo location, as

introduced earlier in the paper.

HI A DYNAMIC NETWORK MODEL OF

GRAIN SILO LOCATION

We now develop a model of where to locate silo

complexes for the storage of grain in relatively large rural,

geographical grain producing areas. As was mentioned in

the introduction, we assume that the grain is transported

from farms, stored temporarily in the silos, and then

transported again to a number of given destinations such

as ports, processing plants, and markets. The model to be

developed has the aim of identifying the location of each

silo to be constructed. Its objective is to minimize the total

cost of the overall grain logistics operation.

Assumptions:

a) A planning horizon T, is given.

b) The sources of grain supply are the centroids

of identified regions.

<) AI grain is transported, at a unit cost, ftom a

region to à silo for intermediate storage and

then to a destination.

d) Regional output

demands are known.


e) Fixed and variable construction costs (based

on capacity) of a silo at each possible

construction site are known.

rates and destination

Db Where each silo is to be constructed, and its

capacity, is to be identified.

g) The shipping schedule from the regions to

the silos, and from the silos to the

destinations is to be determined.

The Objective

The objective is to identify (f) and (g) above that

enable the minimization of total transport and (amortised)

silo construction costs over the planning horizon T.

Notation

Let

Parameters:

T= the number of periods in the planning horizon.

m= the number of regions that are sources of supply.

n = the number of final destinations of the grain

shipments.

p = the number of possible silo locations for intermediate

storage.
R; = the centroid of the iº grain producing region,

i=1,2,..,m.

D; = the j” final destination, j = 1, 2, ...,n.

Ly = the kº possible silo location, k = 1, 2, ...,p.

fi = the fixed cost of silo construction at Ly k = 1,2, ...p.

ve = the variable cost per unit of capacity of silo

construction at Ly, k= 1,2,...,p.

a; = the number of units of supply to be shipped from R,,

i=1,2,...,m.

b; = the number of units of supply to be received by D;,

International Engineering Management Conference 2004

j=1,2,..,n.

ci = the per unit transportation cost from Ri; to Lx,

i=1,2,...,m;k=1,2,...,p.

x = the per unit transportation cost from LtoD,

k=1,2,..,p;j=1,2,..,0n.

cy = the per unit transportation cost from R; to Dj,

i=1,2, ..om;j=1,2,...,n.

M = a sufficiently large number.


2z= the total cost of all transportation and construction.

Decision Variables:

For each period t, t= 1, 2,...,T, let:

Xi = the number of units shipped directly from Ri; to Li.

Xy = the number of units shipped directly fro Ly to D;.

Xj = the number of units shipped indirectly, via at least

one silo, from R;to D;.

ol

w, = the capacity of the silo at Ly, if it is constructed.

O, if a silo is not constructed at Ly,

1, if a silo is constructed at Ly.

The Model: Minimization of transport and total silo

construction costs.

Objective:

. T m P T Pa FP (1)

Min 2=5 5 Desta + DD gt + DU UM)

=) sl dat 151 Rj d=1

Constrainis:

TOP

D LxA

t=] b=l
(The total output of cach Ri is shipped.)

= 12, m. 2)

TP

>, » Xi = bj, j= 12....,n. ()

(The demand of each D; must be met.)

TmTn

» > Xy =5 Dx k=12,..5p. (4

el ia Ta jel

(Material balançe at each L,.)

W=D Xyo K=12.,p. (5)

i=l

(Definition of wr.)

M,, 2 Wi

E=12..,p, (6)

(mi = 0=>y3=0, wy >0 => y; = 1.)

1227
Wy 20, i=12,...,m, (7)

k=1,2,..5p,

(= LZUT

Wy 2 0, k=L2....p, (8)

j=1,2,.,p;

t=1,2,..5,T.

w;>0, k=1,2,...,p.

(9)

y=0,1 k=1,2,.,p.

(10)

The w,'s can be eliminated by:

TOM Ta

We = s z Xi = > > Nigro (11)

t=l in ts ja

k=1,2,.,D.

A more detailed discussion of the issue of grain silo

location has been given by Foulds [21], where various

models, solution techniques and numerical examples are

provided. By taking advantage of the special structure of

the network modelled, numerical instances of practical

dimensions can be solved in reasonable computing time.

For example, the dynamic network flow methods

discussed by Ahuja et al. [19] and by Powell [20], and the

fixed charge model techniques discussed by Knowles [22]

can be used via commercial integer programming codes

such as XPRESS-MP [23].

IV. CONCLUSIONS AND SUMMARY


We have described a dynamic network flow model

for a grain silo location scenario that accommodates

changes in flow over time. Analysis of the model provides

new insights into important aspects of the grain silo

location problem. We believe that the model reported

represents a useful addition to the grain logistics planner”s

toolkit.

REFERENCES

[1] F. L. Hitchcock, “The distribution of a product from several

sources to numerous localities,” | Math and Phys. vol. 20, pp.

224 — 230, 1941.

[2) L. Kantorovich, “On the translocation of masses,” Comptes

rendas (Doklady) de l'academie des sciences de PURSS.

XXXVI 1942.

[3] T. €. Koopmans, “Optimum utilisation of the transportation

system,” Econometrika, vol. XVII, pp. 136 — 146, 1949.

[4] G. Bradley, G. Brown, and G. Graves, “Design and

implementation of large scale prima! transhipment algorithms,”

Mgt Sci, vol, 24, pp. 135, 1997.

1228

[61
E

[8]

[9]

[10]

Eu

[12]

[13]

[4]

[15]

[16]

[17]

118]

[19]

EO]

[21]

BZ

[23]

A. Orden, “The transshipment problem,” Mgt Sci, vol. 2, pp.

276 — 285, 1956.


D. R. Fulkerson, “An out-of -kilter method for solving minimal

cost flow problems,” S/AM J Appl Math, vol. 9, pp. 18 — 27,

1961.

L.R. Ford, and D. R. Fulkerson, Flows in Networks. Princeton,

NJ, USA: Princeton University Press, 1962.

P. A. Jensen, and J, W. Barnes, Network Flow Programming.

New York, NY, USA: Wiley, 1980.

E. W, Dijkstra, “A note on two problems in connection with

graphs,” Numerische Mathematik, vol. 1, pp. 269 — 271, 1959.

Y. 1, Yen, “On Hu's decomposition algorithm for shortest paths

in a network,” Op Res, vol. 19, pp. 983 — 985, 1971.

J. Edmonds, and K. M. Karp, “Theoretical improvements in

algorithms efficiency for network flow problems,” JACM, vol.

19, pp. 248 — 264, 1972,

G. B. Dantzig, Linear Programming and Extensions. Princeton,

NJ, USA: Princeton University Press, 1963.

G. B. Dantzig, and D. R. Fulkerson, “Minimizing the number of

tankers to meet a fixed schedule,” NLRQ, vol. 1, pp, 217-222,

1954.

L. R. Ford, and D. R, Fulkerson, “Maximal flow through a

network,” Can J Math, vol. 8, pp. 399-404, 1956.

D. Gale, “Transient flows in networks,” Mick Marh, vol. 6, pp.

59-63, 1959.
T. Magnanti, and R. Simpson, “Transportation network analysis

and decomposition methods,” Report No, DOT-TSC-RSPD-78-

6, Department of Transportation, Washington, DC, USA, 1978.

L. Chalmet, R. Francis, and P. Saunders, “Network models for

building evacuation,” Mgt Sci, vol. 28, pp. 86 —105, 1982.

R. Vermuganti, M. Oblak, and A. Aggarwal, “Network models

for fleet management,” Dec Sci, vol. 20, pp. 182-197, 1989.

R. Ahuja, T. Magnanti, and J. Orlin, Network Flows: Theory,

Algorithms, and Applicatians. New York, USA: Prentice Hall,

1993.

W. B. Powell, “Network Routing,” in M. O. Ball, T. Magnanti,

€. L. Monma, and G. L. Nemhauser, Eds. Amsterdam, Holland:

Elsevier, 1995, ch. 3,

L. R. Foulds, “Dynamic Network Flow Modeis of Grain Silo

Location and Freeway Lane Direction Specification,” Research

Report Series No. 2004-03, Department of Management

Systems, University of Waikato, Hamilton, New Zealand, 2004.

T. W. Knowles, Management Science. Homewood, Illinois:

Irwin, 1989, pp. 462- 468.

XPRESS-MP, “Dash Optimization,” Blisworth NN73BX, UK.

International Engineering Management Conference 2004

You might also like