Volume 5 Issue 4 Article 5

Artificial Bee Colony Algorithm for Solving Fuzzy Multi-Objective Bed

Allocation Model
Abdulhakeem Luqman Hasan
Directorate of Educational Nineveh Province, Iraqi Ministry of Education, [email protected]

Follow this and additional works at: https://kijoms.uokerbala.edu.iq/home

Part of the Biology Commons, Chemistry Commons, Computer Sciences Commons, Operational Research
Commons, and the Physics Commons

Recommended Citation
Hasan, Abdulhakeem Luqman (2019) "Artificial Bee Colony Algorithm for Solving Fuzzy Multi-Objective Bed Allocation
Model," Karbala International Journal of Modern Science: Vol. 5 : Iss. 4 , Article 5.
Available at: https://doi.org/10.33640/2405-609X.1154

This Research Paper is brought to you for free and open access
by Karbala International Journal of Modern Science. It has been
accepted for inclusion in Karbala International Journal of
Modern Science by an authorized editor of Karbala International
Journal of Modern Science. For more information, please
contact [email protected].
Artificial Bee Colony Algorithm for Solving Fuzzy Multi-Objective Bed Allocation
Model
Abstract
With the improvement of the medical services frameworks rivalry, hospitals face more and more
challenges. In the interim, allotment of resource has a crucial influence on performing competitive
benefits in a hospitals. To choose the suitable beds number is one of the most essential tasks in hospital
administration. Anyway, in true condition, bed allotment choice is a multiple-side problem with weakness
and haphazardness of the information available. It is so sophisticated. Therefore, the research about bed
allotment difficulty is comparatively rare under considering multiple departments, nursing hours, and
stochastic information about arrival and service of patients. In this paper, we improve a fuzzy multi-
objective bed allotment pattern for defeating doubtfulness and various sections. Fuzzy objectives, and
weights are at the same time used to assist the administrators to choose the appropriate beds about
various departments. The suggested pattern uses Artificial Bee Colony (ABC), which is so efficient
algorithm. The research portrays a use of the pattern in a public hospital in Iraq. The outcome has shown
presented an appropriate system for bed allotment and ideal usage.

Keywords
Bed Allocation Problem; Fuzzy Logic; Artificial Bee Colony; Multi-Objective Optimization.

Creative Commons License

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0
License.

This research paper is available in Karbala International Journal of Modern Science: https://kijoms.uokerbala.edu.iq/
home/vol5/iss4/5
1. Introduction frequency distribution connected with a hospital care
units census [5e7], present an analytic access
Busy hospital systems continually provide new dependent upon an adjusted form of the Holt-Winters
challenges to their managers and decision-makers multiplicative quality of being seasonal predicting
because of high demands for service, high costs, pattern.
limited budget and healthcare resources. Therefore, Small number of studies, anyway, regard the dif-
decision-makers are continuously studying produc- ficulty of bed allocation as a problem with many
tivity and competence of present hospital systems objectives [8,9]. Present by a logical computer-
and must be able to evaluate the outcomes of any imitation pattern that: there isn't controlling answer to
changes they make to these systems [1]. Hospitals the difficulty of bed-allocation, so showing the type of
today are encountered with several pressures the difficulty which has many objectives [10e13].
as raising equipment prices, a lack of qualified Describe a GP pattern to examine the act of a Medical
healthcare specialists and restricted hospital facil- Assessment Unit (MAU) and look for answers for
ities [2]. allotment of bed to patients with least possible post-
Generally, bed allocation problem involves the ponement. It is hard to examine the systems act for
fixed number of beds assigned to the different best work flow because of the deterministic nature of
medical and surgical specialties in a hospital. A the model built. Hence [14,15], present a simulation
proper bed allocation is important for cost-efficiency model which consider the factors in length of stay,
or productiveness of hospitals. Few beds appointed number of beds, nurses and doctors in the MAU.
to a specialty may lead the customer to face units Afterwards a GP model is applied to perform
with unsuitable equipment and insufficient staff exchanging analysis, using the results from the
training. This would cause lower quality of care. imitation model. To help lowering postponements and
Also, many beds increase cost by the lack of usage of increase the movement of patients, the hourly allot-
resources [1]. ment plans of resources can be spread out within the
Some studies concentrate on the bed allocation MAU.
problem across various sections, medical specialties, Recently, the bed allocation problem (BAP) has got
or patients' type in a hospital. To save time and get the expanding considerations. Types of mathematical
cost in an easy way [3], develop a mathematical models for BAP were created in the literature,
model used in inventing admission policy for different including the benefits of queuing hypothesis are very
types of patients [1]. Provide access to beds reallo- common. In Refs. [2,16] the model of M/PH/n queues
cations of services from time to time to reduce the was recommended, where M symbolizes patient
expected increase. Their requirement predicting sys- reaching under the influence of Poisson distribution
tem utilizes an M/G/∞ model of queuing for rounding (Markov arrivals), PH marks patient periods of stay
up total number of patient dynamics for every labor. (LoS) under the influence of period-type distribution,
The study accentuates the importance of predictions and n is the number of beds.
must create the foundation for analysis and receive Lack in the number of beds or incorrect bed dis-
conjecture by models of queuing for the aim of pro- tribution can increase waiting times, patients' being in
cedural plainness. the wrong place and not trusting bed management
[4] Regarding the difficulty of allotting a beds [16,17]. Planning beds in an accurate way is impor-
number to various medical and surgical specialties in tant to satisfy patients’ needs, arrange departments
a hospital, taking into consideration that the making a and develop the provided service as to quality and
timetable of medical routines changes above a week amount.
and that the requirement for various medical labors This paper presents the using of Artificial Bee
reveals quality of being seasonal. To make good de- Colony (ABC) which is an effective one in some
cisions concerning the size of each unit, a period public hospitals in Iraq to choose the suitable bed
sequence pattern is improved by using continual numbers and bed allocation taking into consideration
census information. Correspondingly, to decide the number of departments, entire quantity of beds in the

https://doi.org/10.33640/2405-609X.1154
2405-609X/© 2019 University of Kerbala. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/
by-nc-nd/4.0/).
 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244 237

hospital, arrival times, service times and nursing membership of x to A, and it is called the membership
hours. function of A.
This research introduces a fuzzy determination
Definition 2. When the (crisp) set of elements
which is has many objectives and helping pattern
belong to the fuzzy set A, the degree of its membership
dependent upon theory of queuing for allotment of
function goes beyond the level a: Aa ¼ [x 2 X j
beds in a general hospital. An assumption is that beds
mA  a].
number influences the act of a section in the sense of
(1) the approval chance when a recent patient reaches Definition 3. (x) ¼ (f1,…, fm) are the objective functions
and (2) the work hours of nursing. and G(x) are the system constraints. f *i ðxÞ, i ¼ 1 …, is
The objective of this research is to choose the the optimal goal value to the objective.
suitable bed numbers and bed allocation which is a
Definition 4. (fuzzy decision). A fuzzy decision is
necessary task in hospital administration by using
defined in a parallelism to non-fuzzy conditions “as
Artificial Bee Colony (ABC) especially in the general
the choice of activities which at the same time
hospitals in Iraq to raise patient's permission ratio,
fulfill objective functions and restrictions.” In fuzzy
nursing hours and service level.
set theory the intersection of sets usually matches
The paper is organized as follows: Section (2)
to the logical “and “The “decision” in a fuzzy
present a brief revision of fuzzy set theory and
environment can therefore be observed as the
artificial bee colony. Section (3) present the pro-
intersection of fuzzy restrictions and fuzzy objec-
posed fuzzy model of bed allocation problem. The
tive functions [19]. The objectives and restrictions
artificial bee colony to solve FMOBAP has been
are normally not equally important and have
proffered in Section (4). On Section (5), the results
different weights in the multi objective fuzzy de-
and debate are added. Section (6) demonstrate the
cision [20]. In many cases, the decision-maker
differences between fuzzy and non-fuzzy solutions.
cannot exactly determine his relative weights.
The inference of this study has been given in Section
Therefore, the weights are supposed to be fuzzy
(7).
numbers with either triangular or trapezoidal
membership functions.
2. Fuzzy set theory and artificial bee colony
algorithm
2.2. Artificial bee colony
2.1. Fuzzy set theory
The artificial bee colony (ABC) algorithm, which is
Fuzzy set theory (FST), initially recommended a multitude based on collecting of information
by Zadeh [18], is one of the best devices to manage maximum production algorithm, was presented in
the inaccuracy or ambiguity. Under many condi- (2007) by Karaboga for numerical duty maximum
tions, brittle data are insufficient to represent real production by imitating the forging behavior of bee
world circumstances, because human opinions are colonies [21]. The head steps of ABC algorithm can be
usually ambiguous and cannot be guessed by an depicted as explained below.
accurate numeral value. To deal with ambiguity of
human idea, Zadeh [18] before all else announced  Initialization.
the fuzzy logic theory, which was directed to the  Repeat.
sensibleness of doubtfulness caused by imprecision  Occupied bee phase: appoint the occupied bees on
or ambiguity. Hence a great contribution of FST the origins of food in the memory.
(fuzzy set theory) is its skill or ability for sym-  Observer bee phase: appoint the observer bees on
bolizing ambiguity. the origins of food in the memory.
The fundamental definitions and notations below  Seeker bee phase: dispatch the seeker bees to the
will be used in every part of this paper. investigation area for finding out recent food
conditions.
Definition 1. Let X be a universe of discussion; A is
 Up to the time which (requirements are
a fuzzy subset of X if, for all x 2 X, there is a
fulfilled).
number mA 2 [0, 1] appointed to stand for the
238 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244

In the artificial bee colony algorithm, the colony 3. The fuzzy model of BAP
comprises of three sorts of honey bees: occupied
bees, observer bees and seeker bees. One part of the 3.1. A multiobjective BAP model
colony is occupied bees, and the other part is
observer bees. The occupied bees examine the In this study, the model contains two objective
nourishment origin and produce the data of the functions and three constraints. The sets of objectives
nourishment origin to the observer bees. The and constraints are given below.
observer bees select a nourishment origin to take According to Gorunescu et al. [2], the BAP model is
advantage of dependent upon the information defined as a M/PH/n queue. In the hospital there are n
received from the occupied bees. The seeker bee that beds and D departments. For each department i,
belongs to the occupied bees which nourishment i ¼ 1,2, …, D. ni is a symbol for beds number, li is a
origin are rejected discovers a new nourishment symbol for the Poisson reaching ratio, and mi is a
origin accidently. The place of a nourishment origin symbol for the average of length of stay which is
is a reasonable answer to the maximum production estimated by the Phase-type allotment. We conclude
difficulty. Symbolize the nourishment origin number that the average arrivals number through the period of
as SN, the location of the ith nourishment origin as xi stay is li mi . Therefore. In Eq. (5) the probability Pi that
for each i ¼ 1, …, SN, which is a D dimensional some appearances are wasted due to ni beds are filled
vector [14,22]. depending on Erlang's loss formula.
In the ABC algorithm, the ith fitness value i fit for a  n  X ni  
ðli mi Þi ðli mi Þj
reduction difficulty is specified as [23]: pi ¼ ð5Þ
8 ni ! j¼0
j!
>
< 1 if fi  0
fi ti ¼ 1 þ fi ð1Þ where phases of the service time number is defined as j.
>
:
1 þ abcðfi Þ if fi < 0 Therefore, we infer the entrance ratio of patient in
department i as.
Where fi is the cost value of the ith solution. The pos- Pai ¼ 1  Pi ð6Þ
sibility that food origin being chosen by an observer bee
is given by: The total nursing hours in department i, can be
fi ti represented by Eq. (7):
pi ¼ PSN ð2Þ
Fni ¼ tsi  ni ð7Þ
i¼1 fi ti

An applicant answer from the previous one can be where tsi is nursing hours for each bed in department i.
created as: The pattern restriction set (8)e(10) is displayed under.
  The constraint (8) makes certain that the entire beds
vij ¼ xij þ Fij xij  xkj ð3Þ
number (N) is equal to the amount of beds in every
department. The second constraint makes certain that
where k2f1; 2; …; SNg, ksi /and j2 f1; 2; …; Dg
the entire nursing hours between H and L, where H is
are accidentally chosen points out, Fij 2½  1; 1 is
highest hour of nursing with most additional time and L
a without variation distributed accidental number.
is lowest hour of nursing with the least additional time.
The applicant answer is contrasted by the previous
The ultimate constraint makes certain that there is no
one, and the superior answer must be kept [14]. In
department without beds.
case of rejected nourishment origin is xi, the seeker
bee takes advantage a recent nourishment origin X
D
ni ¼ N ð8Þ
depending on:
  i¼1
xij ¼ xmin;j þ randð0; 1Þ xmax;j  xmin;j ð4Þ
X
D
L Fni  H ð9Þ
where xmax;j and xmin;j are the higher and less advanced i¼1
limits of the jth measure of the difficulty's investigation
space [23]. ni > 0; i ¼ 1; 2; …; D ð10Þ
 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244 239

3.2. A fuzzy multiobjective BAP model fi ¼ 1  PaTPa*i ; i ¼ 1; 2; …; D

Pa

Suppose that the administrators want to arrive some fi ¼ tsi  ni TFn*i ; i ¼ 1; 2; …; D

Fn
ambition degree which is not explicitly defined instead of
actually intending to reach the actual maximum or min- s.t
imum of the objective. Objective that makes the resolu- X
D
tion turn into fuzzy programming. The corresponding not ni ¼ N
explicit one objective model for the FMOBAP. That i¼1

procedure sequence, which has been adapted depending X

D
on BAP case, is depicted as follows. L Fni  H
i¼1
Step 1. Build the fuzzy model of BAP invention diffi-
culty depending on the criteria and restrictions of the ni > 0; i ¼ 1; 2; 3; …; D ð11Þ
administrators.
where Pafi and Fn
fi are ambition level that a admin-
Step 2. Decide the less advanced limit and higher limit istrators want to fulfill. The symbol “T” in the
of ambition degree (DM's aim) for each purpose. objective function shows fuzzy inequality which has
Step 3. For the objective functions and fuzzy con- the lingual explanation “essentially bigger than or
straints, find the fellowship or total number of members. equal”.

Step 4. Express in formula the corresponding brittle Step 2. Suppose that ðPa0i ; Fn0i Þ and ðPa1i ; Fn1i Þ, are
model of the fuzzy optimization. the values of the objective function ðPai , Fni Þ
such that 0 or 1 is the degree of fellowship
Step 5. Find the answer to the brittle example by using function, relatively. Those values can be earned by
algorithm of ABC to find the perfect or best answer either a multi-objective solution as one objective
n*.Concerning the previous steps, fuzzy multi-objective problem or a recovery from historical data (DM's
bed allocation problem formula (FMOBAP) can be experience).
depicted as follows.
Step 3. The fellowship function for Pai and Fni are
Step 1. If the administrators or decision makers (DM) described as Eq. (12), Eq. (13) respectively, using non-
has a fuzzy aim such as the objective function should be increasing linear function
considerably equal to or less than some ambition de-
gree, then the FMOBAP can be expressed in a formula
as expressed in the coming section.
find n*i ¼ ½n1 ; n2 ; n3 ; …; nD 

8 0; if PaðnÞ  Pa1i
>
>
>
>
< Pa1i  Pai ðnÞ
mPai ðnÞ ¼ 1 ; if Pa0i  PaðnÞ  Pa1i ; i ¼ 1; 2; …; D ð12Þ
>
> Pa1i  Pa0i
>
>
:
1; if Pai ðnÞ  Pa0i

8 0; if FnðnÞ  Fn1i
>
>
>
>
< Fn1i  Fni ðnÞ
mFni ðnÞ ¼ 1 ; if Fn0i  FnðnÞ  Fn1i ; i ¼ 1; 2; …; D ð13Þ
>
> Fn1i  Fn0i
>
>
:
1; if Fni ðnÞ  Fn0i
240 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244

Step 4. Taking into account that the fuzzy aims and 4.2. Encoding
fuzzy restrictions are dealt with equivalently which
suggests that all determination criteria have identical For FMOBAP, only one parameter is involved
importance. The equivalent brittle one objective pro- which represent the number of beds ni in each
gramming for Eq. (11) can be depicted as explained in department i. For a BAP with D departments, each
the following: food source is encoded as [n1, …, ni, …, nD].
max l ¼ minfmPai ðnÞ; mFni ðnÞg
4.3. Initial population
Pa1  Pai ðnÞ
mPai ðnÞ  i 1
Pai  Pa0i The ABC algorithm begins the investigation by
creating a population of applicant explanations. In our
Fn1i  Fni ðnÞ implementation, this population is randomly generated
mFni ðnÞ 
Fn1i  Fn0i according to the uniform distributions. First of all, we
generate a random value (Ri) for each department i
X
D
ni ¼ N from uniform distribution U [0,1]. Then calculating the
i¼1 number of beds in each department by using Eq. (15),
Eq. (16) respectively.
X
D " # !
L Fni  H Ri
i¼1 ni ¼ round PD  N ; i ¼ 1; 2; …; D  1 ð15Þ
j¼1 Rj
ni > 0; i ¼ 1; 2; 3; …; D ð14Þ
X
D1
ni ¼ N  nj ; i ¼ D ð16Þ
j¼1
Step 5. The final step is to find a solution to the
equivalent brittle model by using ABC algorithm (see
Section (4)). 4.4. Chromosomes evaluation

The aim of Eq. (12) is to make most of the weighted

4. Using ABC algorithm for solving FMOBAP additive of fellowship role of objectives. Assuming
DM's have equal importance, the suitability function
In this study, the BAP has two objectives, namely for ABC may be recorded as the following.
patient admission rate and nursing hours, to optimize.
Main activities that describe our altered ABC are functionz ¼ fðnÞ
illustrated as follows. mPa ðnÞ ¼ minfmPai ðnÞg

4.1. Input information mFn ðnÞ ¼ minfmFni ðnÞg

gðnÞ ¼ …
For the fuzzy multi-objective bed allocation prob-
lem, the software implemented takes into consideration z ¼ minðmPa ðnÞ; mFn ðnÞ þ gðnÞ
as much parameters as possible found in actual hospital
environments: end.

 aspiration level for every objective function. 4.5. Search mechanism

 number of departments D.
 entire quantity of beds in the hospital N. In algorithm of ABC, the occupied bee phase stands
 arrival times (l). for the investigation algorithm capability, and the
 service times (m). observer bee phase stands for the selfish use capability
 nursing hours (ts). of the algorithm.
 The upper limit (H) and lower limit (L) of the In employee bee stage, each bee Xi generates a new
nursing hours. candidate solution Vi in the neighborhood of its present
position as equation below:
 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244 241

  
Vij ¼ round Xij þ aij XKj  Xij ð17Þ Table 1
Arrival rate, the mean of LOS and nursing hours for each bed.
Where i,k 2{1,2, …, SN} is an accidental chosen index I Department li mi tsi (h)
which is different from j2f1; 2; …; Dg is an accidental 1 General Surgery 1 18.115 10.3462 1.2
chosen index, aij 2(0,1) is without variation distributed 2 General Surgery 2 18.192 10.3846 1.15
accidental number. 3 General Surgery 3 18.538 12.156 1.25
4 Urology Surgery 19.153 9.8077 0.65
After all employed bees finish the investigation 5 Fascia Surgery 4.461 2.1923 0.9
process; they have in common the knowledge of 6 Orthopedics 7.961 3.5385 0.55
their food origins with the observer bees through 7 ENT 18 11.1538 0.6
swinging dances. An observer bee appraises the
nectar information taken from all employed bees and ambition level should be logical values to keep away
selects a food origin with a likelihood connected to from unpractical solution.
its nectar quantity. This probable choice is really a Having substituted the parameter values into Eq.
gambling choice mechanism which is depicted as (14). The fuzzy multi-objective model for the BAP is
Eq. (2). acquired as explained in the coming section:
The onlooker bee stage includes adding and sub- max l ¼ minfmPai ðnÞ; mFni ðnÞg
tracting one bed to different random locations in the
source food. Pa1i  Pai ðnÞ
mPai ðnÞ  1 
If a viewpoint cannot be made better over a defined Pa1i  Pa0i
beforehand number (called limit) of cycles, then the
food origin is rejected. Suppose that the rejected origin Fn1i  Fni ðnÞ
mFni ðnÞ  1 
is Xi, and then the scout bee finds out a new food origin Fn1i  Fn0i
to be substituted using Eq. (15), Eq. (16) relatively.
X
7
The function of Scout bee is to keep the difference of ni ¼ 202
the population so that to hinder too fast convergence of i¼1
the algorithm.
X
7
120  Fni  230
4.6. Stopping conditions i¼1

There are no universal stopping conditions accepted ni > 0; i ¼ 1; 2; 3; …; 7 ð18Þ

for ABC algorithm. In this paper, we clearly stop our
algorithm after a given number of repetitions (Ng). As the population size is set as 100, 100 solutions
are generated by the ABC algorithm to find the best
5. Results and discussions solution. The degree fulfillment and attainment level of
aims for the optimum explanation are acquired as
In this section, a real life bed allocation problem is shown below.
presented. For confidential reason, the name of the mPa1 ðnÞ ¼ 0.66, mPa2 ðnÞ ¼ 0.64, mPa3 ðnÞ ¼ 0.28.
hospital concerned is not mentioned. The hospital has mPa4 ðnÞ ¼ 0.45, mPa5 ðnÞ ¼ 0.97, mPa6 ðnÞ ¼ 1,
7 departments and 202 beds. The maximum and min- mPa7 ðnÞ ¼ 0.41.
imum hour of nursing with overtime are 200 h and mFn1 ðnÞ ¼ 0.93, mFn2 ðnÞ ¼ 0.87, mFn3 ðnÞ ¼ 0.76.
150 h respectively. The Poisson arrival rate, the mean
of length of stay and nursing hours for each bed are
shown in Table 1. The parameters for ABC algorithm Table 2
The fewer and higher limit of ambition level.
after many experiments were a limit parameter equal to
30; the number of iterations Ng ¼ 500; the number of i Pa0i Pa1i Fn0i Fn1i
population is 100. 1 43% 100% 20 50
To evaluate, the FMOBAP model will be used to 2 41% 100% 20 50
3 42% 100% 19 50
find a solution to the hospital problem. Firstly, the
4 43% 100% 15 30
fewer and higher limit of ambition level for all pur- 5 67% 100% 4 20
poses must be clarified. The values of Z 0i and Z 1i are 6 72% 100% 5 20
shown in Table 2. It is essential to recognize that the 7 40% 100% 15 30
242 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244

mFn4 ðnÞ ¼ 0.47, mFn5 ðnÞ ¼ 0:14, mFn6 ðnÞ ¼ 0.14, Table 3
mFn7 ðnÞ ¼ 0.36 (see Fig. 1). Comparison between the new and previous allocation decisions.
We compare our solution with the earlier allotment Dept. Previous New
determination of their hospital. As revealed in Table 3, N Pa Fn N Pa Fn
n is the beds number allotted for each department, Pa 1 32 66.34% 38.40 40 80.68% 48.00
is the entrance permission ratio, and Fn is the nursing 2 32 64.36% 36.80 40 78.57% 46.00
hours. In addition, Fig. 2, explains the alterations in act 3 32 54.80% 40.00 34 58.08% 42.50
of the patient entrance permission ratio in every 4 32 64.68% 20.80 34 68.40% 22.10
5 10 99.98% 9.00 7 99.10% 6.30
department, at the same time Fig. 3, shows the alter- 6 32 100% 17.60 13 100% 7.15
ations in act of the nursing hours in every department. 7 32 60.94% 19.20 34 64.51% 20.40
The benefits of the allotment determination created by Total 202 73.01% 181.80 202 78.48% 192.45
ABC are summed up as follows.
The fuzzy model and ABC algorithm balance the
6. Comparison fuzzy and non-fuzzy solution
resource rivalry between various departments. The
acting productivity and degree of administration in the
The bed allocation objective function and its re-
hospital are improved. In the earlier allotment deter-
strictions can be define as:
mination, section 6 has the greatest entrance permis-
sion ratio about 100% and section 3 has the least X7 X
7 
entrance permission ratio about 54.80%. In compari- min z ¼  ci * Pai ðnÞ þ ci * Fni ðnÞ ð19Þ
i¼1 i¼1
son, in the new allotment the greatest entrance
permission ratio is 100% in department 6 and the X
7
Pai ðnÞ X 7
Fni ðnÞ
lowest admission rate is 58.08% in section 7. There- min z ¼  ci * max
þ ci * max ð20Þ
i¼1
Pai i¼1
Fni
fore, in the new allotment, determinations, not only the
entrance permission ratios are stable, but also the least Where, Pamax max
i and Fni is the highest importance of
entrance permission ratios are notably made better. parallel targets which are prized from pre-processing
Almost, resemble outcomes are also informed in Step (warm-up time) in ABC algorithm series.
nursing hours. As can be noticed, the greatest nursing As an ultimate outcome, the comparison of act
hours is 40 h in the previous situation, 48 in the new criteria between fuzzy and non-fuzzy explanation for
situation. the given hospital scenario is presented by Table 4. The
From an extensive perspective, the general patient fuzzy solution comparing with non-fuzzy solution
entrance permission ratio and nursing hours are grown showed an improvement in admission rate of patient
from 73.01% to 181.8 to 78.48% and 192.45 and nursing hours from 75.45% to 186.95 h to 78.48%
correspondingly. and 192.45 h, relatively. It appears to be that FMOBAP

Fig. 1. Membership function for fuzzy objectives.

 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244 243

NEW PREVIOUS

1.00%

1.00%
1.00%
0.99%
0.81%

0.79%

0.68%
0.66%

0.65%

0.65%
0.64%

0.61%
0.58%
1 2 0.55%
3 4 5 6 7

Fig. 2. Comparisons of patient admission rates.

NEW PREVIOUS
0.48

0.46

0.43
0.40
0.38

0.37

0.22
0.21

0.20
0.19
0.18
0.09

0.07
0.06

1 2 3 4 5 6 7

Fig. 3. Comparisons of nursing hours.

Table 4
The comparison between fuzzy and non-fuzzy solution.
Dept. Non-Fuzzy Sol. Fuzzy Sol.
7. Conclusions
N Pa Fn N Pa Fn
This paper proposes a fuzzy multiobjective decision
1 30 62.48% 36 40 80.68% 48.00 aiding model based on queuing theory for BAP in a
2 36 71.68% 41.4 40 78.57% 46.00
3 37 62.94% 46.25 34 58.08% 42.50
hospital. The Artificial Bee Colony (ABC) is used to
4 33 66.55% 21.45 34 68.40% 22.10 optimize the proposed model. Solved by ABC, the
5 11 100% 9.9 7 99.10% 6.30 beds in 202-general hospital in Iraq, is reappropriated.
6 21 100% 11.55 13 100% 7.15 Imitation results explain that resource rivalry between
7 34 64.51% 20.4 34 64.51% 20.40 various departments is more stable and the general
Total 202 75.45% 186.95 202 78.48% 192.45
patient entrance permission ratio and nursing hours in
the hospital are both raised. Therefore, the level of
model can efficiently directed when the beds should be
service and the resource use in the hospital are made
exchanged, what quantity it is demanded and at which
better at the same time.
department must be added.
244 A.L. Hasan / Karbala International Journal of Modern Science 5 (2019) 236e244

References [12] S.M. Ballard, M.E. Kuhl, The use of simulation to determine
maximum capacity in the surgical suite operating room, in:
[1] E.P.C. Kao, G.G. Tung, Bed allocation in a public health care Proceedings - Winter Simulation Conference 7, 2006,
delivery system, Manag. Sci. 27 (5) (1981) 507e520. pp. 433e438, https://doi.org/10.1109/WSC.2006.323112.
[2] F. Gorunescu, S.I. McClean, P.H. Millard, A queuing model for [13] H.R. Feili, Improving the health care systems performance by
bed-occupancy management and planning of hospitals, J. Oper. simulation optimization, J. Math. Comput. Sci. 7 (2013) 73e79,
Res. Soc. 53 (I) (2002) 19e24, https://doi.org/10.1057/pal- https://doi.org/10.22436/jmcs.07.01.08.
grave/jors/2601244. [14] J. Oddoye, M. Yaghoobi, M. Tamiz, D. Jones, P. Schmidt, A
[3] A. Esogbue, A. Singh, A stochastic model for an optimal pri- multi-objective model determine efficient resource levels in a
ority bed distribution problem in a hospital ward, Oper. Res. 24 medical assessment unit, J. Oper. Res. Soc. 58 (2007)
(5) (1976) 884e898, https://doi.org/10.1287/opre.24.5.884. 1563e1573.
[4] S. Lapierre, D. Goldsman, R. Cochran, J. Dubow, Bed alloca- [15] J. Nguyen, P. Six, D. Antonioli, P. Glemain, G. Potel,
tion techniques based on census data, Soc. Econ. Plan Sci. 33 P. Lombrail, P. Le Beux, A simple method to optimize hospital
(1999) 25e38. beds capacity, Int. J. Med. Inform. 74 (I) (2005) 39e49.
[5] M. Cote, A note on bed allocation techniques based on census [16] F. Gorunescu, S.I. McClean, P.H. Millard, Using a queueing
data, Soc. Econ. Plan Sci. 39 (2005) 183e192, https://doi.org/ model to help plan bed allocation in a department of geriatric
10.1016/j.seps.2004.01.006. medicine, Health Care Manag. Sci. 5 (4) (2002) 307e312,
[6] G. Chryssolouris, K. Dicke, M. Lee, On the resources allocation https://doi.org/10.1023/A:1020342509099.
problem, Int. J. Prod. Res. 30 (2007) 2773e2795, https:// [17] M.D. Muynck, H. Bruneel, S. Wittevrongel, Analysis of a
doi.org/10.1080/00207549208948190. discrete-time queue with general service demands and phase-
[7] X.D. Li, P. Beullens, D. Jones, M. Tamiz, Optimal Bed Allo- type service capacities, J. Ind. Manag. Optim. 13 (2017)
cation in Hospitals, Lecture Notes in Economics and Mathe- 1901e1926, https://doi.org/10.3934/jimo.2017024.
matical Systems, vol. 618, Springer, 2009, pp. 253e265. [18] L.A. Zadeh, Fuzzy sets, Inf. Control 8 (1965) 338e353.
[8] S.C. Kim, I. Horowitz, K.K. Young, T.A. Buckley, Flexible bed [19] H.J. Zimmermann, Fuzzy programming and linear program-
allocation and performance in the intensive care unit, J. Oper. ming with several objective functions, Fuzzy Sets Syst. 1 (1978)
Manag. 18 (4) (2000) 427e443. 45e55.
[9] L. Luo, L. Li, X. Xu, W. Shen, L. Xaio, A data-driven hybrid [20] H.J. Zimmermann, Fuzzy Set Theory and its Applications,
three-stage framework for hospital bed allocation: a case study Kluwer Academic Publisher Group, Dordrecht, 1991.
in a large tertiary hospital in China, computational and math, [21] D. Karboga, B. Basturg, Powerful and efficient algorithm for
Methods Med. (2019) 1e12, https://doi.org/10.1155/2019/ numerical function optimization: artificial bee colony (ABC)
7370231. algorithm, J. Glob. Optim. 39 (2007) 459e471.
[10] J. Oddoye, M. Tamiz, D. Jones, P. Schmidt, A Simulation [22] A. Coello, B. Lamon, A. Van Veldhuisen, Evolutionary Algo-
Model for Health Planning in a Medical Assessment Unit with rithms for Solving Multi-Objective Problems, Springer, 2007,
Multi-Objective Output Analysis, University of Portsmouth, https://doi.org/10.1007/978-0-387-36797-2.
UK, 2007. Technical report, Management Mathematics Group. [23] K. Deb, B. Basturk, On the performance of artificial bee colony
[11] Z.F. Antmen, S.N. Ogulata, The capacity planning of intensive (ABC) algorithm, Appl. Soft Comput. 8 (2008) 687e697,
care units via simulation: a case study in university hospital, Int. https://doi.org/10.1016/j.asoc.2007.05.007.
J. Appl. Math. Stat. 51 (2013) 214e235.