AMERICAN JOURNAL OF SOCIAL AND MANAGEMENT SCIENCES

ISSN Print: 2156-1540, ISSN Online: 2151-1559, doi:10.5251/ajsms.2012.3.1.1.7

2012, ScienceHu, http://www.scihub.org/AJSMS

Comparative analysis of linear and multi-objective model application in a

private hospital healthcare planning in Nigeria
Salami, A. O.
Department of Business Enterprise Management, College of Management Science, Federal
University of Agriculture, Abeokuta, Nigeria
ABSTRACT
Goal Programming is designed to provide decision makers with an opportunity to satisfy diversified goals
while linear programming is designed to optimize an objective function subject to several constraints. These
techniques, when used properly can effectively aid decision makers in short and long range strategic
planning. The objective of this study is thus to compare the solution methodologies of both the optimization
(Linear) and Multi-Objective model in arriving at policy guided decision making in health care industry. OlaOlu Hospital, a medium size private hospital in Ilorin was used as a case study for the paper. The findings of
the study reveal that Goal Programming (GP) can be used to solve single as well as multiple objective
decision problemss, subject to linear constraint equations. It has also been shown in this study that, in order
to derive a meaningful solution, decision makers must give considerable attention to the formulation of both
the LP and GP model. It is thus concluded that GP is applicable to many settings and types of decisions,
and from a practical perspective, is worthy of a trial application.
Keywords: Healthcare Planning; Optimization; Multi-Objective and Modeling
INTRODUCTION
The Nigerian health care has suffered several down-falls,
despite Nigerian's strategic position in Africa, the country is
greatly underserved in the health care sphere. Health
facilities (health centers, personnel, and medical
equipments) are inadequate in this country, especially in
rural areas (Onwujekwe et. al., 2010). While various
reforms have been put forward by the Nigerian government
to address the wide ranging issues in the health care
system, they are yet to be implemented at the state and
local government area levels. According to the 2009
communiqu of the Nigerian national health conference,
health care system remains weak as evidenced by lack of
coordination, fragmentation of services, dearth of
resources, including drug and supplies, inadequate and
decaying infrastructure, inequity in resource distribution,
and access to care and very deplorable quality of care. The
communiqu further outlined the lack of clarity of roles and
responsibilities among the different levels of government to
have compounded the situation.
Unarguably, problems in the health care system of any
country abound to a certain extent.[Dougherty and
Conway, 2008; Davis, et. al., 2005; Green-Pedersen and
Wilkerson, 2006 and Moe et. al.,] Although health has the
potential to attract considerable political attention, the
amount of attention it actually receives varies from place to
place. In their commentary of the 3T's road map to
transform US health care, Denise Dougherty and Patrick H.
Conway rightly stated a step by step transformation of the
US health care system from 1T 2T 3T which is
required to create and sustain an information-rich and
patient-focused health care system that reliably delivers
high-quality care (Dougherty and Conway, 2008).

Provision of timely information aimed at combating possible

health menace among many other things is an important
function of public health. Hence, inadequate tracking
techniques in the public health sector can lead to huge
health insecurity, and hence endanger national security,
etc. (Moe et. al., 2007; Jones et. al., 2009)
In the complex financial environment of todays healthcare
industry, there is great need for administrators to be
familiar with efficient methods of allocating scarce
resources. Many healthcare facilities [HCF] have become
similar to large business enterprises; they are complex
organizations established to achieve certain objectives
through integration and use of limited resources. One might
think of an HCF as an organization whose primary
production output goal is service. The inputs consist of:
(1) labour of various types (physicians, nurses, aides,
technicians),
(2) fixed assets (buildings, beds, equipment) and
(3) working capital, converted to tangible variable assets
(food, drugs, bedding, supplies).
The services output of an HCF can measured in terms of
quantity and quality. Quantity can be measured in terms of
admissions, discharges, duration of stay, number of
outpatients, number of procedures, etc. Output is more
difficult to measure because of many subjective elements.
However, the quality component of a service is comparable
with the quality component of a product; it is a function of
its input quality. In other words, if an HCF has highly
qualified physicians and nurses, and technologically
advanced equipment, the quality of service received by its
patients is likely to be high.
Various quantitative techniques have been suggested in
the literature that can efficiently optimize the allocation of

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

scarce resources. Linear programming [LP] is considered

to be one of the most powerful quantitative techniques
available. Many readers are familiar with the basic
assumptions and concepts related to LP, a method
successfully utilized over four decades. It deals with solving
management problems of optimization of a single objective,
subject to several linear constraints. Although LP is widely
used in decision-making processes, it has a major limitation
that restricts the users of the technique to narrowing their
problems to a single objective function.

Note, the literature contains conflicting definitions (see

Blake and McCarl; Ignizio [1978,1983]; Romero [1989,
1991]). For example: ( a) goal programming has been used
to refer to multiple objective problems with target levels; (b)
multiobjective programming has been used to refer to only
the class of problems with weighted or unweighted multiple
objectives; (c) vector maximization has been used to refer
to problems in which a vector of multiple objectives are to
be optimized; and d) risk programming has been used to
refer to multiobjective problems in which the objectives
involve income and risk. Multiobjective programming
involves recognition that the decision maker is responding
to multiple objectives.

In real life, decision situations are frequently characterized

by multiple, conflicting objectives. For instance, today
HCFs operates in a competitive environment, which is
heavily influenced by external factors including regulators,
competitors, suppliers, customers, shareholders and the
public. These factors in turn make profit maximization no
longer the sole objective for management to strive to
achieve.

Generally, objectives are conflicting, so that not all

objectives can simultaneously arrive at their optimal levels.
An assumed utility function is used to choose appropriate
solutions. Several fundamentally different utility function
forms have been used in multiobjective models. These may
be divided into three classes: lexicographic, multi-attribute
utility and unknown utility.

Beside profit maximization, management now has many

other objectives to attain, including improved market share,
improved relationships with the public, production of a
quality product, minimized employee unrest, minimized
pollution, and realization of a certain rate of return on
investment. Many of these may conflict; and may compete
for scarce resources. If the existence of multiple conflicting
objectives is recognized, it follows that an ideal decision
technique is one that is able to take them all into account.
The objective of this study is thus to compare the solution
methodologies of both the optimization (Linear) and MultiObjective model in arriving at policy guided decision
making in health care industry.

The lexicographic utility function specification assumes the

decision maker has a strictly ordered preemptive
preference system among objectives with fixed target
levels. For example, a lexicographic system could have its
first priority goal as income of not less than $10,000; the
second priority as leisure of no less than 20 hours a week;
the third as income of no less than $12,000, etc. This
formulation is typical of "goal programming models."
(Charnes and Cooper (1961); Lee, (1972)). The various
goals are dealt with in strict sequential order - higher goals
before lower order goals. Once a goal has been dealt with
(meeting or failing to meet the target level), its satisfaction
remains fixed and the next lower order goal is considered.
Consideration of the lower level goals does not alter the
satisfaction of higher level goals and cannot damage the
higher level goals with respect to target level attainment.

Conceptual Clarification: Optimization of a single

objective oversimplifies the pertinent objective function in
some potential mathematical programming application
situations. Arguments can also be made following
Simon ( 1954) that optimization is not as appropriate as
statisficing. These two statements introduce the general
topic of multiobjective programming. Multiobjective
programming formally permits formulations where: (a)
solutions are generated which are as consistent as
possible with target levels of goals; ( b) solutions are
identified which represent maximum utility across multiple
objectives; or (c) solution sets are developed which contain
all nondominated solutions. Multiple objectives can involve
such considerations as leisure, decreasing marginal utility
of income, risk avoidance, preferences for hired labor, and
satisfaction of desirable, but not obligatory, constraints.

Multi-attribute utility approaches allow tradeoffs between

objectives in the attainment of maximum utility. The most
common form involves maximization of the sum of linearly
weighted objectives. This type of formulation has been
used by Candler and Boeljhe (1977); and Barnett, Blake
and McCarl (1982). The third utility approach involves an
unknown utility function assumption. Here the entire Pareto
efficient (nondominated) solution set is generated so that
every solution is reported wherein one of the multiple
objectives is as satisfied as it possibly can be without
making some other objective worse off (Geoffrion (1968)).

A discussion of this area requires some definitions. An

objective is a measure that one is concerned about when
making a choice among the decision variables (something
to be maximized, minimized or satisfied like leisure, risk,
profits, etc.). A goal implies that a particular goal target
value has been chosen for an objective. We will use
"multiple objective programming" to refer to any
mathematical program involving more than one objective
regardless of whether there are goal target levels involved.

The lexicographic multiple objective formulation is not

precisely a LP problem. It has many structural
characteristics in common with a LP problem; however, a
conventional objective function is not defined, nor can a
single LP formulation reflect imposition of the sequential
ordering of the goals. Rather, an iterative procedure is
needed (Lee, 1972). Essentially, the approach is to solve
problems for each of the goals sequentially. When
considering the ith goal solve the problem .The new
variable wr gives the amount that the goal level (grjXj) is

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

less than the target value (Tr), while glr is the current level
of goal r. When i = 1 the problem minimizes the shortfall
from the first goal target level, subject to the LP constraints.
One of two solution situations will then occur. Either the
optimum value of w1 (denoted w*1) equals 0,

The objective function maximizes multi-dimensional utility

summed across all objectives. Each objective is weighted.
The second equation sums the level of each objective into
the variable glr. The third expresses satisfaction in terms of
the normalizing factor. The fourth represents resource
availability limitations, the fifth expresses nonnegativity
constraints and the sixth allow the goal level to be positive
or negative (note the normalizing factor must be of the
appropriate sign).

indicating full satisfaction of the first goal, or w*1 0,

indicating the goal cannot be fully satisfied. Subsequently,
a second problem is solved. This problem is virtually
identical to the first, except w2 is minimized and a
constraint is appended indicating that w1 cannot be any
worse than the optimum value obtained at the end of the
solution of the first stage (w1*). This requires: 1) if goal 1
was met before, then goal 1 will continue to be met (i.e., w1
must be less than or equal to zero); or 2) if goal 1 was not
met, then the deviation from goal 1 will not get bigger than
the minimum deviation obtained at the previous iteration.
Thus, the prior objective is constrained to be no worse off
than it was before. This problem, in effect, explores
alternative optimums where we hold the prior objectives at
their optimum values, then try to optimize the satisfaction of
the subsequent objectives.

This is again a linear program. The formulation is adapted

from Lee (1972) and is used in Barnett, et al. (1982).
The other approach to multiobjective programming involves
an unknown utility function assumption. Instead, the entire
nondominated set of alternatives is generated. The
formulation for this approach is exactly like the first one
under the weighted tradeoff section above except that all
possible weights are utilized in the problem. This particular
approach has been studied extensively, (see, for example,
the bibliographies in Steuer (1968); and Ignizio, 1983) but
does not appear to be very empirically useful.
A CASE STUDY

This procedure is executed for all R goals where different

deviation variables are minimized at each stage and a
constraint is added holding all previous deviations to
maximum values prohibiting the earlier objectives from
becoming worse off. Lee (1972)
presents a more
comprehensive discussion of the procedure while the
example below gives an empirical application.

Linear Programming Applied to a Healthcare Problem

(Optimization Model): Ola-Olu Hospital is a medium-size
HCF, lomocated at Muritala Muhammed Way, in Ilorin, the
Capital of Kwara State, Nigeria. For the sake of simplicity,
assume that the HCF is specialized in performing four
types of surgery: Tonsillectomies, appendectomies, hernias
and cholecystectomies. The performance of these
surgeries is constrained by three resources: operating
room hours, recovery room bed hours, and surgical service
bed-days. Assume that the director of the HCF would like
to determine the optimum combination of surgical patients
that maximizes the total contribution to profit, given the
information in Table 1.

The second utility function type involves tradeoffs between

various objectives. Such problems can be formulated as
conventional linear programs. There have been two
alternative formulations of this problem. They differ in their
assumptions about target levels. The first formulation
(appearing for example in Candler and Boeljhe, 1977) does
not take into account target levels, maximizing the
weighted sum of the quantities of each objective. This
where cr is the weight which expresses the importance of
the rth objective in the context of the decision maker's total
utility and Nr is a normalizing factor which converts the goal
values so they are valued somewhere around one.. The cr
coefficients would be in utility units per percent deviation
from full satisfaction at the normalizing factor for the rth
objective achieved; glr is the amount of rth objective in the
optimal solution and qr is the proportional satisfaction
amount of rth objective relative to the normalizing factor.

Where
X2
X3
X4
Let:S2
S3

X1
=
=
=
S1
=
=

=the number of tonsillectomy patients

the number of appendectomy patients
the number of hernia patients, and
the number of cholecystectomy
= the idle hours of the operating room,
the idle hours of recovery room, and
the idle days of the surgical service beds.

Table 1: Ola-Olu Hosptial Data

TYPES OF SURVIAL PATIENTS
CAPACITY
X1
X2
X3
X4
Operating room
3
4
8
6
1,100 hours
Recovery room
8
2
4
2
1,400 bed hours
Surgical services
4
6
4
2
400 bed hours
Average contribution to profit
N210
N260
N280
N300
Since the above problem has only one objective function [i.e. the determination of the optimal combination of the surgical patients to
maximize total average contribution], it can be solved by LP. The problem can be readily solved, giving the optimal solution shown in the
final simplex tableaus in Table 2

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

Table 2: Final Simplex Tableau of LP

Objective row Cj
Objective
Variable
Constraint
Column
column
column
280
X3
100
0
S2
900
300
X4
50
Zj
43,000
Cj Zj row
Solution Stub

210
X1

260
X2

280
X3

300
X4

0
S1

0
S2

0
S3

-0.60
7.8
1.3
222
-12

-1
2
2
320
-60
Body

1
0
0
280
0

0
0
1
300
0

0.20
-0.60
-0.10
26
-26

0
1
0
0
0

-0.30
0.40
0.40
36
-36

Table 2 indicates that the optimal combination of the

surgical patients that maximizes the total profit, given the
available resources, is X3 the best way for the HCF to
maximize its contribution to profit is to treat only 100 hernia
patients [X3] and 50 cholecystectomy patients [X4]. This will
earn the healthcare facility N43,000 and it will leave it with
900 idle hours in the recovery room [S 2 = 900]. While on
the surface insisting on these mixes of surgery types may
appear ludicrous, two points need to be made regarding
the management of patient loads:
1) It is possible to defer and/or accelerate the timing
of some surgeries to accommodate, or at least
approach, this type of optimal mix.
2) Where past history forecasts a non-optimal mix, it
is appropriate to manage the reconfiguration of
the scarce resources in such a way that it will
optimally handle the mix that history predicts.
Now assume that Ola-Olu Hospital seeks to
achieve the following multiple objectives:
1) Achievement of at least N50,000 in profit in a
specified period of time, given the available
resources.
2) Minimisation of idle capacity of the available
resources, i.e. operating room hours, recovery
room bed-hours, and surgical services bed-days.
The solution cannot be obtained by LP since there is more
than one objective, but GP is appropriate for handling the
problem. The basic assumptions underlying LP are
identical for GP. However, GP is capable of handling
decision situations which involve multiple objectives. In
fact, the solution to the previous LP problem can also be
obtained through GP.

Similarly, if d is driven to zero, this means that the overachievement of the goal will not be realized.
From the foregoing discussion, it can be deduced that
deviational variables are mutually exclusive. The
+
relationship is mathematically expressed as d x d = 0.
The steps needed to structure a GP model are:
1) Goals are identified and expressed as constraints.
2) Goals are analyzed to determine the correct
+
deviational variable[s] needed for them, d i, d i, or
both
3) A hierarchy of importance among goals is
established by assigning to each of them a preemptive priority factor, Pj. These pre-emptive
represents the highest priority, P2 the second
highest, and so on. The Ps indicates a simple
ordinal ordering of the goals.
Once these above steps are completed, the
problem can be quantified as a GP model as
follows:
Minimise:
m
+
i
Pj [d i + d 1-], where j 1,2,3, , n
i=1
Subject to:
n
+
[aijxj] - d i + d i = bi, where i = 1,2,3, , m
The following are the problem formulation and solution:
i
+
Minimise: OX1 + OX2 + OX3 + OX4 + P0 [d -] + Od 1 + P 2
[d 2 + d 3 + d 4]
Subject to:
i+
210X1 + 260X2 + 280X3 + 300X4 + d 1 + d 1 = N50,000 .
Target Profit [1]
3X1 + 4X2 + 8X3 + 6X4 + d 2 = 1,100 Operating room
hours
[2]
8X1 + 2X2 + 4X3 + 2X4 + d 3 = 1,400 Recovery room
bed-hours
[3]
4X1 + 6X2 + 4X3 + 4X4 + d 4 = 400 Surgical services beddays
[4]
i+
X1, X2, X3, X4 , d 1, d 1 d 2, d 3 and d 4 0
Non-negativity constraint
Note that the first goal equation has two deviational
+
variables [d 1 and d 1] and d 1 appears in the objective
+
function with a P1 coefficient while d 1 has a coefficient of
zero. The reason for this manipulation is that, since goal
one requires making at least N50,000, there is no need to
+
put any restriction on d 1. This is equivalent to the
maximization of profit problem in LP. Also, the insertion of
+
d 1 in equation [1] is required to determine if and by how
much the minimum of N50,000 profit has been exceeded.
+
In other words, variable d 1 will indicated the difference
between the goal profit and actual profit.

The Goal-Programming Model (Multi-Objective Model):

The basic steps used in structuring an LP model are
identical to those for GP. The major difference between the
two models is that the GP model does not optimize
[maximize/minimize] the objective directly, as in the case of
LP. Instead, it attempts to minimize the deviations between
the desired goals and the realized results. Also, these
goals must be prioritized in a hierarchy of importance.
To quantify this prioritization, each goal is expressed as an
equation, and a deviational variable[s] is/are assigned to it.
Deviational variables can be positive or negative. A positive
+
deviation variable [d ] represents over-achievement of the
goal. A negative deviation variable
[d } represents under-achievement of the goal. If the desire
is not to under-achieve the goal, d should be driven to zero.

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

Equation [2], [3] and [4] include only the underachievement variables d 2, d 3 and d 4, respectively, without
inserting the over-achievement variables. This is logical
since the objective of the second goal is to minimize the
idle capacity of all the scarce resources. The above
problem was solved by the simplex method of GP. Table 3
presents the optimal solution [final simplex tableau].

Since the above GP problem has two prioritized goals, both

should have a Cj Zj row, which Table 3 verifies. In LP, if
the desire is to minimize an objective function, the optimal
solution will be reached when all the Cj Zj values under
the columns to the right of the constant column become
zero, or positive [Cj Zj >0]. The point of optimality for the
lower priority goals, such as P2, will be reached when all
the negative values of the Cj Zj row for the P2 in the body
section become zero or positive, provided that the negative
values have corresponding zero values of the Cj Zj row
for P1.
Interpretation of Table 3
The problem solution is to treat 100 hernia patients [X3] and
50 cholecystectomy patients [X4]. This patient mix will bring
to the HCF N43,000 and leave it with 900 idle hours in the
recovery room. This can be noted in the solution stub of
Table 3 [d 3 = 900]

Interpretation of the GP Solution: The solution stub [Cj

Zj] of the simplex tableau of GP is similar to that of LP
except that in GP each goal has its own Cj Zj row, this
due to the ordinal ranking of goals. The highest priority goal
occupies the lowest Cj Zj row, the next Cj Zj row
represents the second priority goal, and so on. The Cj Zj
row will show whether a goal is optimized or sub-optimized
[partially satisfied].

Table 3: Goal Programming Solution

Objective row Cj
0
0
0
0
P1
Objective
Variable
Constraint
X1
X2
column
column
column
P1
d-1
7,000
-12
-60
0
X3
100
-0.6
-1
P2
d-3
900
7.8
2
0
X4
50
1.3
2
Cj Zj
P2
-900
-7.8
-2
Cj Zj
P1
-7,000
12
60
Solution Stub
Body
Source: Solution of the Model using simplex method Algorithm

X3

P2
X4

0
1
0
0
0
0

0
0
0
1
0
0

P2
d-1
1
0
0
0
0
0

P2
d+1

d-2

d-3

d-4

-1
0
0
0
0
1

-26
0.20
-0.60
-0.10
1.6
26

0
0
1
0
0
0

-36
-0.30
0.40
0.40
0.60
36

The solutions in Table 2 and 3 are similar. The reason is

that in LP, the objective was to maximize profit, and the
maximization of profit is a function of the optimal utilization
of the scarce resources. In GP, the prioritized objectives
were: first to achieve at least N50,000 in profit [this is
equivalent to the maximization of profit in LP] and secondly,
to minimize idle capacity of the available resources [this is
equivalent to the optimum utilization of scarce resources in
LP]. Thus, it is not surprising that the solutions in Tables 2
[LP] and III [GP] are similar.
The solution stub of Table 3 also shows that both goals [P1
and P2] are under-achieved. The first priority goal [P1] was
under-achieved by N7,000 and second [P2] was underachieved by 900 hours.
Measuring
Goal
Achievement: Short-term
goal
achievement: Table 3 solution is short term because with
limited resources their utilization could not generate the
minimum desired profit. The questions now become: What
resource[s] should be increased? By how much? What
patient mix should be treated to attain all of the specified
goals [profit and minimization of unused capacity]? Such
questions involve the longer-term planning.
Longer-term goal achievement: In order to optimize all of
the previously mentioned goals [achieve at least N50,000
profit and minimize idle capacities of all the available
resources], the previous GP model can be readjusted. The
short-term equations [2]- [4] can be rearranged to make
them suitable for the longer term, as follows:

Minimise: OX1 + OX2 + OX3 + OX4 + P1 [d 1] + d

+
+
+
+
+ 0d 2 + P2 [d 3] + 0d 3 + P2[d 4] + 0d 4

+
1

+ P2 [d 2]

Subject to:
+
210X1 + 260X2 + 280X3 + 300X4 + d 1 + d 1
=
N50,000
[5]
+
3X1 + 4X2 + 8X3 + 6X4 + d 2 + d 2= 1,100
[6]
+
8X1 + 2X2 + 4X3 + 2X4 + d 3 d 3 = 1,400
[7]
+
4X1 + 6X2 + 4X3 + 4X4 + d 4 d 4= 400
[8]
+
+
+
+
X1, X2, X3, X4 , d 1, d 1 d 2, d 2, d 3, d 3, d 4, d 4 0
non-negativity constraint
Comparing the above equations with those for the short
term, the following differences can be noted:
1) In the longer-term model, an over-achievement
+
variable [d ] has been inserted in each capacity
+
constraint equation [equations [6] [8]. The d
represents capacity expansion.
+
+
+
2) The insertion of d 2, d 3, and the d 4 variables in
the longer term model requires assigning them
zero coefficients in the objective function. This
means that there is no restriction on the
expansion of these capacities as long as this
expansion satisfies the predetermined objectives.
The solution to the above formulation is shown in
Table 4.
Interpretation of the Solution
The solution stub of Table 4 shows that, in order to fully
optimize all the predetermining goals in the longer term, the
following strategy should be adopted:

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

1.

2.

The above solution optimizes both priority goals. The Cj

Zj rows for P1 and P2 in the solution stub verify this point.
The cells at the intersections of both Cj Zj, P1 and Cj Zj,
P2 with the constant column both have zero values. This
means there is no under-achievement value for them.
Table 5 presents a comparison of the short and longer term
solutions.

Increase the surgical service bed-days by 300.

+
This is noted in the d 4 row in the solution stub,
where this variable has a value of 300 in the
constant column.
Treat only tonsillectomy and hernia patients. The
best combination of these patients will be 88
hernia patients [X3] and 130 tonsillectomy patient
[X1]

Table 4: GP Final Simplex Tableau Long Range

Objective row Cj
0
0
0

P1

P2

d4

X3

X4

d1

7300

-5

-3

0.5

-0.5

-1

X3

88

0.5

1.8

0.20

-0.2

-0.1

0.1

+
d1

2,230

-120

-1

-1

26.9

-26.9

16.2

16.2

X1

130

106.
2
-0.2

-0.10

0.1

0.2

-0.2

Cj Zj

P2

15

-0

Cj Zj

P1

+
4

Solution Stub

d3

+
d3

P2

X2

X1

d2

+
d2

P2

Constraint
column

Variable
column

+
d1

Objective
Column

+
4

Body

Source: Solution of the Model using simplex method Algorithm

Table 5: Comparison of the Short and Longer term Solutions of GP
Short-term Solution
Longer-term Solution
Goal[s] optimized : none
Goals optimized: All
Patient mix solution: 100 hernia patients
Patient mix solution: 130 tonsillectomy
[X3] and 50 cholecystectomy patients [X4]

Patients [X1] and 88 hernia patients [X3]

Source: Results in table 3 and 4

have to be couched in the best of managerial thinking

about values: personal, institutional, and
environmental. And solutions do not specify their
intrinsic relationships to hosts of variables not
considered, and the potential consequences of
decision implementation, which may be both
functional and dysfunctional.

CONCLUSION
The foregoing demonstrates the advantages of GP
over LP. Tables 2 and 3 shows that GP can be used
to solve single as well as multiple linear objectives
subject to linear constraint equations. It has also
been shown that, in order to derive a meaningful
solution, decision makers must give considerable
attention to the formulation of a GP model. They must
be precise and accurate in determining overachievement and under-achievement variables
needed for every goal, for lack of accuracy can yield
a long-term solution when the intent is a short-term
solution.

The process of building or designing a model such as

goal
programming
model
is
not
always
straightforward. Although, such model will invoke
certain simplifications and assumptions that make the
problem we are dealing with analogous to some other
situation for which there is a ready-made model and
solution. Oladipo and Salami, [2004]. The goal
programming model used in this write-up is
applicable to all types of organization. However, this
study has been designed to help managers in the
Healthcare sector think through the practice of goal
programming technique as a tool of optimizing profit
and other major organizational objectives. Although,
attention of the researchers was focused on Ola-Olu

But let the reader beware. GP is not a panacea for a

multitude of problems seeking solutions. Optimal
solutions may not achieve primary goals. Some of the
underlying assumptions may be found to be relatively
unrealistic; for example, many relationships are really
curvilinear. Solutions require implementations
which may be impractical or difficult. Decisions still

Am. J. Soc. Mgmt. Sci., 2012, 3(1): 1-7

hospital, Ilorin, the study can also be extended to

other manufacturing and/or service organizations.

Denmark and the US. Journal Eur Publ Pol. 13: 103952

From the foregoing, it is recommended that decision

makers or policy makers should exploit the unified
approach, model and philosophy provided by the
multi-objective mathematical programming model so
as to better solve their administrative and other
similar problems.

Ignizio, J. P. (1983): Generalized Goal Programming: An

Overview. Computers and Operations Research 10:
277-89
Jones, D. S., Tshimanga, M., Woelk, G. Nsubuga, P.,
Sunderland, N. L., Hader, S. L. (2009): Increasing
leadership capacity for HIV/ AIDS programmes by
strengthening public health epidemiology and
management training in Zimbabwe. Human Resources
for Health. 7: 69

REFERENCES
Barnett, D., Blake and McCarl, B. A. (1982): Goal
Programming via Multidimensional Scaling Applied to
Sengalese Subsistence Farms. American Journal of
Agricultural Economics. 64 : 720-27

Lee, S. M. [1972], Goal Programming for Decision

Analysis, Auerbach, Philadelphia
Moe, J. L., Pappas, G., Murray, A. (2007): Transformational
Leadership, transnational culture and political
competence in globalizing health care services: A case
study of Jordans King Hussein Cancer Center.
Globalization and Health. 3: 11

Blake, B. F. and McCarl, B. A. (1983): Goal Programming

via Multidimensional Scaling Applied to Sengalese
Subsistence Farming: A Reply. American Journal of
Agricultural Economics. 64 : 823-33.
Candler, W and Boehlje, M. (1977): Use of Linear
Programming in Capital Budgeting with Multiple Goals
American Journal of Agricultural Economics. 53 : 325330

Nigeria National Health Conference 2009 Communique

Abuja, Nigeria Available from http://www.ngnhc.org
Onwujekwe, O., Onoka, C., Uguru, N., Nnenna, T.,
Uzochukwu, B., and Eze, S., (2010): Preferences for
benefit packages for community-based health
insurance: An exploratory study in Nigeria.

Davis, K., Schoenbaum, S. C., Audet, A. M. (2005): 2020

vision of patient-centered primary care. Journal of
General International Medication. 20: 953-7

Romero, C. (1991): Handbook of Criteria Issues in Goal

Programming. Oxford, New York: Pergamon Press.

Dougherty, D., and Conway, P. H (2008): Road Map to

Transform US health care. The How of high-quality
care. JAMA: 299: 2319-21

Simon, H. A. (1954): Models of Man. New York: John Wiley

and Sons.

Green-Pedersen, C., Wilkerson, J. (2006): How agendasetting attributes shape politics: Basic dilemmas,
problem attention and health politics developments in