Counterfeit Resistant Framework for Single Machine Booking and Grouping in a Production Network Planning Issue

Manikandan Abiraman. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 7, Issue 2, ( Part -1) February 2017, pp.60-66
www.ijera.com DOI: 10.9790/9622 -0702016066 60 | P a g e
Counterfeit Resistant Framework for Single Machine Booking
and Grouping in a Production Network Planning Issue
Manikandan Abiraman
Department Of Mechanical Engineering, Panimalar Engineering College, Chennai, Tamil Nadu.
ABSTRACT
This paper addresses a generation and outbound dissemination planning issue in which a few occupations must
be prepared on a solitary machine for conveyance to clients or to different machines for further handling. We
accept that there is an adequate number of vehicles. Likewise, it is expected that the conveyance cost is free of
bunch size, yet it is reliant on every excursion. In this paper, we introduce an Artificial Immune System (AIS)
for this issue. The goal is to limit the total of the aggregate weighted number of late employments and the cluster
conveyance costs. A cluster setup time must be included before preparing the principal work in every group.
Utilizing computational test, we contrast our technique and a current strategy for the specified iss ue in writing,
specifically Simulated Annealing (SA). Computational tests demonstrate the huge change of the AIS over the
SA.
I. INTRODUCTION
Two key operational capacities in an
inventory network planning are generation and
conveyance operations. In a production network, it
is basic to incorporate or all the while consider
these two capacities and plan and calendar them
together in a planned way to accomplish an ideal
operational execution. Established booking issues
did not consider circulation and conveyance cost,
so considering both the conveyance cost and
planning objective in incorporated model is an
essential subject. Chen investigated the Integrated
Production and Outbound Distribution Scheduling,
in particular IPODS, models and arranged these
issues into five gatherings. Issues with a target
capacity that consider both the machine planning
and conveyance are fairly intricate. In any case,
they are more useful. In any case, the collection of
writing on joined improvement group conveyance
issues, particularly with huge size arrangement, is
little. Lobby and Potts concentrated the issue of
creation booking on a solitary machine under the
bunch accessibility presumption (dissemination
planning) with a few destinations including the
aggregate of stream times, most extreme delay, and
the quantity generally occupations. Cluster
accessibility supposition implies that every one of
the employments shaping a group get to be
distinctly accessible for later preparing or dispatch
just when the whole clump has been handled . They
displayed dynamic-programming calculations for
limiting the specified goals with conveyance costs
when the groups are to be conveyed to a few clients
independently. This paper addresses the limiting
whole of aggregate weighted number of late
occupations and conveyance costs for multi-client
in a solitary machine environment and presents an
AIS calculation for comprehending it surprisingly.
The base number of late occupations, i.e., is
acquired by the polynomial Moore's calculation for
the single machine environment. The weighted
form of issue, i.e., 1, is hard. For An, a Fully
Polynomial Time .Guess Scheme, FPTAS, was
given by Sahni . Later, Gens and Levner enhanced
it twice. Furthermore, Hallah and Bulfin created
Branch and Bound, B&B, technique for this issue
considering zero prepared time and non-zero
prepared time .Hochbaum and Landy proposed a
dynamic programming calculation for the clumping
form of the issue, i.e., , in which occupations are
prepared in groups which require setup time , and
later Brucker and Kovalyov enhanced it. By the by,
none of these reviews considered the conveyance
costs. Steiner and Zhang tended to the comparable
issue, i.e., booking and clumping issue conveyance
to a client, considering the limiting aggregate of the
aggregate weighted number of late employments
and conveyance costs on the single machine with
cluster setup time; they introduced ideal properties
and a pseudopolynomial time DP calculation for
the ideal arrangement. Additionally, they
introduced a pseudopolynomial DP and a FPTAS
for confined instance of multicustomer, where late
employments are conveyed independently toward
the finish of calendar. As of late, Assarzadegan and
Rasti-Barzoki have concentrated the issue of
limiting the most extreme lateness, due date task,
and conveyance costs on a solitary machine. They
introduced two scientific programming models and
two metaheuristic calculations, a versatile
hereditary calculation, and a parallel-mimicked
strengthening calculation, for settling it. A few
scientists have connected metaheuristic
calculations to take care of planning issue.
RESEARCH ARTICLE OPEN ACCESS

In this exploration, we exhibit an AIS
technique for this issue and contrast it and SA and
MINLP approaches presented by Mazdeh et al.
utilizing computational test is representation of our
issues with respect to documentation that Chen
presented for these sorts of issues. This
documentation implies that there is the single
machine for handling occupations with group setup
time, and adequate vehicle by boundless limit and
coordinating conveyance technique for sending the
bunches to clients. Coordinating conveyance
technique implies that requests are transmitted to
every client without the steering issue is the
aggregate weighted number of late employments
and is the aggregate conveyance costs where and
are conveyance costs unit and the quantity of
clumps for every client, individually. Chen
displayed a vital audit on the writing with respect
to coordinated generation and dissemination
planning models. Thus, we don't broadly expound.
Whatever remains of this paper is sorted
out as takes after: Section 2 contains the issue
definition. AIS structure and our proposed
calculations are given in area 3. Area 4 portrays
and breaks down the computational outcomes.
Furthermore, the last area contains our decisions.
II. NOTATIONS AND PROBLEM
DEFINITION
2-1. Notations
Indexes
Job index
Customer index
Parameters:
Number of jobs
Number of customers
Batch setup time for jobs belong to customer
Processing time of job
Due date of job Weight of job
Delivery costs for sending batch to customer
Decision variables:
One if job be tardy and zero otherwise
Number of batches for customer Immune system
notations:
Number of iteration in local search Population size
Local optimum factor in iteration
Mutation rate
Affinity value of antibody
Control factor of decay
2-2. Problem definition
There are clients and one producer in
which every client orders employments to the
maker and is the quantity of occupations. No
occupation can be acquired. Every occupation has
an imperative coefficient. It is expected that
employments require one operation and maker does
it by a solitary machine, with preparing time. The
due date of occupation is. Occupations are prepared
and sent in groups to every client. A clump can
contain employments just for a similar client. This
suspicion is regular in writing for instance observe.
We accept that every cluster has a sequence
independent-group setup time. There is an adequate
number of vehicles, and the conveyance cost is
autonomous of bunch size, and it is appeared by for
client for each trip. The quantity of groups for
client is spoken to by which is a choice variable.
The goal is to limit the total of the aggregate
weighted number of late employments and
conveyance costs. In this way, as specified prior,
this issue can be appeared by 1.
III. ARTIFICIAL IMMUNE SYSTEM
The resistant framework is a data handling
and self-learning framework that offers motivation
to outline AIS. In the most recent decade, the
invulnerable framework has drawn huge
consideration as a potential wellspring of
motivation for novel ways to deal with take care of
complex computational issues. A few scientists
utilized the AIS to tackle the booking class issues.
In this paper, a metaheuristic calculation in light of
AIS interestingly is utilized to limit the aggregate
weighted number of late employments and
conveyance costs in two-level inventory network.
There are a few insusceptible calculations, for
example, negative choice calculation, clonal
determination calculation, and manufactured
resistant systems. In this paper, arrangement
method depends on the clonal choice calculation, in
which just the most elevated fondness antibodies
multiply. So as to comprehend the AIS, some
preparatory natural terms are required to be
portrayed.
Immune cells: B-cells and T-cells are the two
fundamental gatherings of invulnerable cells. These
cells help perceive a practically boundless scope of
against genic examples.
Antigens (Ag): These are sickness creating
components that are partitioned into two sorts of
antigens: self and non-self. Non-self-antigens are
sickness bringing on components, while self-
antigens are innocuous to the body.
Anti bodies (Ab): It is a particle delivered by a B-
cell because of an antigen and has the specific
property of joining particularly with the antigen,
which incited its arrangement.
In the organic procedure, when an antigen
contacts with the insusceptible framework, it
discharges an arrangement of B-cells, introduce in
the immunological memory, with the capacity of
recognizing and disposing of the antigens. Those

B-cells that perceive the antigens with a negligible
partiality are decided for cloning and the quantity
of clones of a specific cell is characterized by its
antigen proclivity. The cells experience substantial
hypermutation after the cloning procedure with a
specific end goal to attempt to kill the antigen. The
cloning and change procedures are rehashed until
the antigen is wiped out. At last, the cells with the
most astounding proclivity are incorporated into
the immunological memory. Hypermutation and
receptor altering are two essential attributes of the
safe framework. They help in the development of
the descendants, as antibodies present in memory
reaction must have a higher liking than those in the
prior essential reaction. Hypermutation is like the
transformation administrator of the hereditary
calculation. The distinction lies in the rate of
adjustment that relies on upon the antigenic
proclivity.
When all is said in done, the antibodies
with lower antigenic liking are hypermutated at a
higher rate when contrasted with the antibodies
with higher antigenic fondness. This marvel is
known as receptor altering, which administers the
hypermutation. The primary errand of
hypermutation is to direct toward nearby ideal,
though receptor altering gets away from the
neighborhood optima.
In whatever is left of this segment, our AIS
properties are presented in detail.
3-1.Encoding schema
In the proposed calculation, a counter
acting agent incorporates a few qualities, with the
end goal that every quality demonstrates the bunch
number of every occupation. This encoding plan is
appeared in Figure 1. This Figure demonstrates that
occupation 1 puts in bunch 5, work 2 puts in clump
2, and alternate employments in the comparative
way. As specified some time recently, proposed
recreated tempering calculation to tackle this issue.
In their calculation, arrangements are encoded by a
grid delineated in Figure 2 where the lines speak to
the bunches and the segments speak to the clients.
For example, if the component in line 2 and
segment 1 is one, the main request of client 1 is
appointed to clump 2. Along these lines, for every
arrangement, a network is framed with
components, though base on our encoding, for
every arrangement, a cluster is shaped with
components.
Jobs: 1 2 3 4 5 6 7 8 9 10
Batches: 5 2 2 4 1 3 5 2 1 3
3-2. Affinity Calculation
In the proposed calculation, it is expected
to ascertain liking of antibodies. Since the objective
is to limit the target work and the proclivity esteem
ought to be augmented in the AIS calculations, the
less target work esteem is considered as the liking
esteem
3-3. The proposed algorithm:
The main framework of the proposed algorithm is
described as follows:
1. Initialization.
2. While (has not met stop criterion) do
3. Local search.
4. Proliferation.
5. Hypermutation.
6. New generation.
7. End.
I. Initialization
In this stage, an irregular beginning
populace of size is made. For every counter acting
agent, the estimation of every quality is resolved
haphazardly in the range 1, in which the estimation
of every quality is exceptional. In this paper, with
the assistance of starting examination, the extent of
populace ( ) is viewed as equivalent to 12.
II. Local search
For every counter acting agent, the accompanying
procedure is done circumstances:
One quality is chosen haphazardly. At that
point, the estimation of this quality that speaks to
the related group is changed to another cluster that
incorporates either no employment or alternate
occupations of the client of that quality;
subsequently, the new arrangement is shaped. At
that point, the fondness estimation of the new
arrangement is computed. In the event that the
blown condition is fulfilled, the counter acting
agent will be supplanted by the new arrangement.
___________(1)
Where is the partiality estimation of
immune response, is the proclivity estimation of
the new arrangement, and is the neighborhood ideal
calculate the emphasis. This variable prompts to
escape calculation from neighborhood ideal. At to
start with, esteem is equivalent to zero; when the
best arrangement is not enhanced in three back to
back cycles, surprisingly, its esteem will be
equivalent to beginning quality. In every emphasis,
esteem is diminished base on condition (2) as takes
after: __________________(2)
Where is the neighborhood ideal calculate
emphasis 1, is the nearby ideal figure cycle , α is
the underlying estimation of neighborhood ideal

variable, is the quantity of cycle, and is the cycle
that the estimation of neighborhood ideal element
will be equivalent to introductory incentive
surprisingly.
With this condition, the estimation of
nearby ideal calculate the last cycle will be
equivalent to zero. Subsequently, the assorted
qualities in the essential cycles is more prominent
than the last emphasess. In this paper, with the
assistance of introductory examination, the
estimations of and are viewed as equivalent to
0.005 and 80, individually.
III. Multiplication
In this procedure, a few clones are created
from every counter acting agent. As in Reisi and
Moslehi, the accompanying condition is utilized to
compute the quantity of clones that every immune
response produces._______________(3)
Where is the quantity of clones, is the
measure of the underlying populace, and is the
aggregate likelihood of the counter acting agent.
For every counter acting agent, is gotten by
isolating its fondness esteemby the whole of all the
immunizer affinities.
IV. Hypermutation
After the expansion organize, the change
administrator is performed for every clone. In
transformation method, one quality is chosen
arbitrarily. At that point, the estimation of this
quality that speaks to its cluster is changed to
another clump that incorporates either no
employment or alternate occupations of the client
of that quality. the transformation rate of every
clone is figured in light of the accompanying
condition:
exp (4)
V. New populace
After the hypermutation procedure is done
and the liking of the hypermutated arrangements
are computed, select the settled number of best
immune response for the people to come. In this
paper, in light of beginning investigation, is viewed
as equivalent to 40% of PS
VI. Stop basis
Utilizing computational pre-test, the stop
basis is considered as takes after: if the best
arrangement is not enhanced after five back to back
emphasess or after an aggregate number of 200
cycles, the calculation will be halted
IV. COMPUTATIONAL RESULTS
In this area, with a specific end goal to
assess the execution of AIS, both the little and
medium-measure issues are considered. The AIS
and SA calculations were coded utilizing Matlab
2009 and keep running on a PC with a 2.93 GHz
CPU and a 2.00 GB RAM. The MINLP model was
coded in GAMS and illuminated by BONMIN
solver, on the grounds that our pre-test
demonstrates BONMIN is the most effective solver
for tackling the said issue. In little and medium-
measure issues, we have contrasted aftereffects of
the proposed calculation and MINLP and SA,
individually. The points of interest will be given in
the accompanying related subsections.
4-1. Issues with,
The quantity of employments in little
estimated issues was set 4,7,10. The quantity of
clients for each was characterized by a uniform
conveyance in the interim 1,. Preparing times,
group setup times, and occupation weights were
haphazardly created whole numbers from the
uniform dissemination characterized on 1100, 0
0.5̅, and 1 100, individually. In light of the cluster
conveyance costs values, we produced two classes
of issues, in particular An and B, for every given
number of the employment. For class An and class
B, the interims that the conveyance expenses were
produced haphazardly are 0 and 0 10,
individually. For every class, we produced three
subclasses, to be specific 1, 2, and 3, in light of the
due dates values; along these lines, we have six
gatherings, in particular A-1, B-1, A-2, B-2, A-3,
B-3. For gatherings (A-1, B-1), (A-2, B-2), and (A-
3, B-3), the interims that the due dates qualities
were produced haphazardly are 0 0.5̅ ̅ , 0 ̅ ̅ and 0
5̅ ̅ , separately.
For every employment number in every
gathering, 10 issues were created arbitrarily.
Subsequently, absolutely 180 (3*2*3*10) issues in
little measured issues were being created and
unraveled. A 300-second time requirement was
considered, and if the issue couldn't be understood
with respect to this limitation, then the technique
would never again be utilized for that issue. The
aftereffects of the analysis for little measured issues
are appeared in Table 1. Segment "Number of the
arrangement in which" of Table 1 demonstrates that
for all issues, AIS has delivered target work, i.e.,
add up to cost, less or equivalent to MINLP show.
In detail, AIS has delivered target esteem the same
as MINLP for 67.22% of issues and completely a
superior outcome for 32.78% of issues.
In issues with four occupations, both
MINLP model and AIS calculation have found the
ideal answer for all issues in every gathering,
except the normal run time of AIS calculation is
littler than the normal run time of MINLP model
for every gathering of four jobs. The normal run
time of issues with 4 is 0.38 second and 32.26
seconds for AIS and MINLP demonstrate,

individually. In issues with 7 and 10 occupations,
the normal of deviation in A-3 is littler than A-2
and is littler than A-1 in A-2. This implies as the
due dates diminish the distinction amongst AIS and
MINLP, target work increments. What's more, the
normal of deviation in B-1 is littler than A-1; in B-
2, is littler than A-2; in B-3, is littler than A-3. In
this way, as conveyance costs diminish the contrast
amongst AIS and MINLP, target work increments.
Tab. 1. The result of experiment for small-sized problems, comparing AIS with MINLP
Delivery Due Date Number of the solution in which Ave. of CPU time (s) Costs
C classes Subclass AIS<MINLP *
AIS=MINLP MINLP<AIS MINLP AIS
A
4
1
2
3
0
0
0
10
10
10
0
0
0
48.04
41.87
7.67
0.39
0.37
0.37
B 1
2
0
0
10
10
0
0
30.78
50.45
0.36
0.36
Cost
*AIS<MINLP implies that AIS has a superior outcome (less aggregate cost) than MINLP
-
Since a few issues have not been illuminated inside 300 seconds by GAMS, the st. of CPU time couldn't figured
for them
4-2. Issues with
In this area, we have thought about
consequences of our proposed calculation, i.e.,
AIS, with SA proposed by in medium-sized issues.
The quantity of employments in medium-sized
issues was set 50,80,110,140. All parameters were
produced like the past, however the quantity of
issues for every occupation number in every
gathering was set 20; consequently, absolutely 480
(4*2*3*20) issues in medium-sized issues were
created. Table 2 demonstrates the aftereffect of the
computational test for this issues.
Tab. 2. The result of experiment in medium-sized problems, comparing AIS with SA
Delivery Costs Due Date Number of the solution in which Ave. of CPU time (s) (%) (%)
C classes Subclass AIS<MINLP *
AIS=MINLP MINLP<AIS MINLP AIS
7
10
A
B
A
B
3
1
2
3
1
2
3
1
2
3
1
2
3
0
8
6
0
3
4
0
9
10
1
9
8
1
10
2
4
10
7
6
10
1
0
9
1
2
9
0
0
0
0
0
0
0
0
0
0
0
0
0
14.76
-
-
-
-
-
0.41
-
-
-
-
-
-
Classes AIS<SA AIS=SA SA<AIS SA AIS Avg. max Avg. max
50 A
B
1
2
3
1
2
15
20
14
18
16
0
0
6
0
0
5
0
0
2
4
1.500
1.652
1.193
3.548
2.840
5.875
4.875
3.749
4.639
5.451
3.3568
13.767
51.982
2.738
1.992
18.676
58.762
475.000
10.109
10.634
0.674
0.000
0.000
0.041
0.043
5.877
0.000
0.000
0.768
0.412
3 14 6 0 2.402 3.123 16.253 156.690 0.000 0.000
80 A
B
1
2
3
1
2
20
19
20
20
20
0
0
0
0
0
0
1
0
0
0
2.645
2.564
1.917
4.939
4.428
12.279
10.788
5.756
12.007
12.196
8.976
39.657
35.094
6.011
8.645
15.738
109.876
215.165
28.028
61.892
0
0.089
0
0
0
0
1.772
0
0
0
3 19 1 0 4.190 7.401 24.106 155.767 0 0

Subclass
Delivery Costs Due Date Number of the solution in which Ave. of CPU time (s) (%) (%)
Subclass
Classes AIS<SA AIS=SA SA<AIS SA AIS Avg. max Avg. max
3 19 1 0 3.963 17.211 108.856 595.625 0 0
B 1
2
20
20
0
0
0
0
9.941
10.794
39.466
31.236
19.169
28.379
93.312
89.023
0
0
0
0
3 19 1 0 6.759 17.689 42.434 187.731 0 0
Table 2 demonstrates that the AIS
calculation has found a superior arrangement, less
target work than SA calculation in 450 (93.75%)
issues, and its target capacity is equivalent to SA
for 18 (3.75%) issues; consequently, AIS has
fathomed 97.5% of all. problems with less equal
total cost with respect to SA; SA has presented a
better solution for only 2.5% of problems.
However, the average run time of AIS is larger
than SA. Column shows the deviation of SA from
AIS when AIS has presented the better result than
SA; Column shows the deviation of AIS from SA
when SA has presented the better result than AIS.
The average deviation of SA from AIS for 93.75%
of problems, for which AIS has presented better
result than SA, is 26.04%, while the average
deviation of AIS from SA for 2.5% of problems
that SA has presented better result than AIS, is
0.20%. The maximum deviation for SA and AIS is
595.63% and 5.69% respectively. These results
shows that AIS is more efficient than SA. It is
obvious from Table 2 that problems in class B has
more average run time, for both SA and AIS, than
problems in class A. So, as delivery costs increase,
more time was required until the stopping criteria
hold. In general, the value for subclass 3 is greater
than subclass 2, and for subclass 2 is greater than
subclass 1. Therefore, as due dates increase, the
deviation of SA fromAIS increases.
V. CONCLUSION
This paper presents an AIS algorithm for
the scheduling and batching a set of jobs on a
single machine with batch setup time for delivery
to customers. In order to evaluate the performance
of the AIS algorithm, computational tests are used.
The computational results show that the proposed
AIS framework is more efficient than the MINLP
and the SA proposed by . Considering some
constraints such as the number of vehicle and
capacity for each vehicle, other machine
configurations for a manufacturer, such as the
parallel machine or flow shop, routing delivery
method, instead of directing delivery method, can
be suggested for future works. In addition, another
function for the total costs, such as total weighted
lateness and delivery costs are suggested as well.
REFERENCE
[1]. Agarwal R, Tiwari M, Mukherjee S.
Artificial immune system based approach
for solving resource constraint project
scheduling problem, The International
Journal of Advanced Manufacturing
Technology, Vol. 34, Nos. 5-6, (2007), pp.
584-593.
[2]. Gao J. A novel artificial immune system
for solving multi objective scheduling
problems subject to special process
constraint, Computers & Industrial
Engineering, Vol. 58, No. 4, (2010), pp.
602-609.
[3]. Engin O, Döyen A. A new approach to
solve hybrid flow shop scheduling
problems by artificial immune system,
Future Generation Computer Systems,
Vol. 20, No. 6, (2004), pp. 1083-1095.
[4]. Chandrasekaran M, et al. Solving job shop
scheduling problems using artificial
immune system, The International Journal
of Advanced Manufacturing Technology,
Vol. 31, Nos. 5-6, (2006), pp. 580-593.
[5]. Steiner G, Zhang R. Approximation
algorithms for minimizing the total
weighted number of late jobs with late
deliveries in two-level supply chains,
Journal of Scheduling, Vol. 12, No. 6,
(2009), pp. 565574
[6]. Dasgupta D, Yu S, Nino F. Recent
advances in artificial immune systems:
models and applications, Applied Soft
110 A
B
1
2
3
1
2
20
20
18
20
20
0
0
2
0
0
0
0
0
0
0
4.022
3.810
2.895
7.781
7.302
22.207
15.892
9.891
22.657
20.461
14.758
30.030
50.717
10.256
13.377
42.654
95.338
300.000
44.755
75.304
0
0
0
0
0
0
0
0
0
0
3 19 1 0 5.488 11.302 34.832 218.156 0 0
140 A 1
2
20
20
0
0
0
0
5.037
5.250
34.630
26.626
14.564
45.090
72.348
88.867
0
0
0
0

Computing, Vol. 11, No. 2, (2011), pp.
1574-1587.
[7]. Steiner G, Zhang R. Minimizing the
weighted number of late jobs with batch
setup times and delivery costs on a single
machine, Multiprocessor Scheduling,
(2007), pp. 85-98.
[8]. M’Hallah R, Bulfin R. Minimizing the
weighted number of tardy jobs on a single
machine with release dates, European
Journal of Operational Research, Vol. 176,
No. 2, (2007), pp. 727-744.

Counterfeit Resistant Framework for Single Machine Booking and Grouping in a Production Network Planning Issue

More Related Content

Viewers also liked (20)

Similar to Counterfeit Resistant Framework for Single Machine Booking and Grouping in a Production Network Planning Issue (20)

Recently uploaded (20)

Counterfeit Resistant Framework for Single Machine Booking and Grouping in a Production Network Planning Issue