Integration Aggregate Production Planning and Maintenance Using Mixed

Integer Linear Programming

M. Erfanian1, M. Pirayesh 2
1,2
Industrial Engineering Department, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
([email protected], [email protected])

Abstract - Aggregate production planning is a planning planning are considered and the kind of maintenance
process and control of entire production activities to meet policy is time based.
the variable demands of aggregate products. The goal is to In the introduced model in this article, the failures
determine the production and inventory levels to minimize and system capacity reduction rate are directly entered to
the total production cost with limited resources of workforce the production planning model and the costs of preventive
and equipment. On the other hand, the aim of maintenance
planning is to improve the efficiency of equipment to satisfy
maintenance planning are correlated with the production
the production requirements. In this paper an integrated costs. The cost of preventive maintenance planning is a
model of aggregate production planning and maintenance function of different factors, viz. the lost production
planning is developed to determine the optimal plan of opportunity, the direct cost of preventive maintenance
production and preventive maintenance in each period. To planning and tools and as well, the unemployment cost of
exhibit the performance of the introduced model, a case production workforce during the time of preventive
study is conducted in a pharmaceutical company and its planning. On the other hand, the cost of conduction of
results are expressed. emergency maintenance planning can be a function of
different factors. Such as percentage of product
Keywords – Aggregate production planning, Emergency
production wastes during production due to system failure
maintenance, Preventive maintenance
and also the direct production costs and tools and even the
unemployment cost of the production forces during the
maintenance time.
I. INTRODUCTION
In this article, an integrated production and
The aim of aggregate product planning is the best use maintenance planning model is introduced to minimize
of workforce and production equipment. The word the production system costs and reach an optimal solution
aggregate means that a planning in a specific level is for balancing the production and maintenance planning in
integrated and the entire products that use the same the midterm level.
resources and equipment are considered as a single The issue of production and maintenance planning
product. In the aggregate production planning the integration in the production systems due to importance
production and inventory levels, is determined in a has been accounted as a special status in the researches.
midterm planning horizon from 3 to 8 months [1]. Budai et al. [2] carried out a complete review of these
One of the assumptions that is present in the issues and classified them into 4 categories including
traditional aggregate production planning models is that conceptual models, economic production quantity,
the production system is always available and it works production system models with the buffer inventory and
with maximum capacity. Nowadays considering the short integrated models. The introduced model in this article
products’ shelf life, most of the mass production systems nestles in the integrated group models, and for this reason
have been converted into multi-product production only this group of articles has been assessed. Schmidt [3]
systems. Even the increase of product variety leads to the conducted a complete review on the integrated models.
complexity of the production equipment. The equipment Weinstein and Chung [4] introduced a three section model
of such systems in order to avail profit should always be to assess the maintenance policy of an organization
located in the maximum operation power. For this reason, wherein an aggregate production planning model by a
when a defect or halt occurs in them, it will intensely linear programming model created in the first stage, and
influence the exploitation of production system. Thus, in the second stage. A master production scheduling for
having an efficient maintenance planning is an essential the minimization of the weighted deviation from the aims
affair for production companies. Both maintenance and of first stage was prepared. Even in the third stage the
production planning occupy the machine available need of working centers via capacity planning using the
capacity, so the relationship between maintenance and specification of resources for depiction of equipment
production is established through machine capacity. failure during the planning period has been determined.
In the operation, three kinds of general policy exists Aghezzaf et al. [5] analyzed a production system that was
for the maintenance planning including the emergency exposed to the random failures. In this article, it was
maintenance planning, preventive maintenance planning, proposed that the actions carried out for the maintenance
and condition-based monitoring. The accomplishment of planning on the production system leads to the reduction
each of the mentioned maintenance plans needs to spend of the production system capacity. Aghezzaf and Najib [6]
some time and capacity of production system. In this introduced a production planning and preventive
article the emergency and preventive maintenance maintenance integration model. In this model they have

978-1-5090-3665-3/16/$31.00 ©2016 IEEE 927

 Proceedings of the 2016 IEEE IEEM

converted a production system to a few independent wherein the occurrence rate of these failures follows an
production lines wherein each of the lines are susceptible exponential distribution with specified mean.
to failures and that leads to the loss of a production The decisions related to planning include decision
capacity during the maintenance period. For the states in about the production level of each product in the regular
which the preventive maintenance planning is conducted and overtime, the shortage level of each product in each
in periodical form, they presented a model of non-linear period as well reduction or increase rate at the workforce
mixed integer and solved it by the heuristic methods. level of production section. On the other hand, the
Besides for the preventive maintenance planning that is decisions related to the maintenance planning are those
conducted in the non-periodical form also introduced a for all of the defined failures on each system to decide
mixed linear integer model. Zhao and Wang [7] which kind of policy is used. In fact, it responds to this
introduced a production planning and maintenance query that which of the failures is in preventive form and
planning integrated model with an assumption that the which one should be implemented in an emergency form.
production capacity will be lost due to preventive The introduced model in this article is a multi-period
maintenance planning or emergency maintenance and multi-product model. The shortage in this model is
planning. They also considered the reliability of permissible and it is from the backorder type. The
machines’ and used an iterative solving approach to find a planning horizon is considered in the monthly manner,
solution for their non-linear model. Yildirim and Nezami and in each month it could have different working days
[8] presented a production planning and maintenance compared to the other months. The time of preventive
planning model with this assumption that the processing maintenance planning is at the beginning of each planning
periods of products is influenced by the system functional period and the time of corrective or emergency
conditions. Gehan, et al. [9] presented another integrated maintenance planning is immediately after occurrence of
model for production and maintenance planning. In this each failure.
article, it was assumed that the production systems Among the important assumptions in this model is
confronted to wear out due to effect of production and that the quality of preventive maintenance planning
over time and this wear out could lead to the probability implementation will be relatively higher than the
increase of the failure during the production period. emergency maintenance planning. In this model it is
In the earlier articles, the production planning is assumed that if the preventive maintenance planning in
discussed without consideration of the entire aspects the initiation of a period is not conducted for one of the
related to planning. For example, in these articles the failure factors, definitely the related failure during the
discussion of workforce resources planning, and the same period will occur with the normal occurrence rate
increase and reduction of costs in the level of workforce and the average number of occurrences of that failure is
resources are considered as one of the most important achieved according to the exponential distribution in the
production resources. In addition, the shortage cost of duration of each period. The implementation of
products and also the production capacity were calculated preventive maintenance planning for each of the failure
during the regular production periods and overtime causes that the related failure occurrence rate reduces to
without separation. In these articles, the effect of affairs zero for a specified time period. After the passage of that
related to maintenance planning on the production system time period again the failure occurrence rate returns back
costs is not specified; and no correlation is established to the previous states and the average number of the
between them viz. the research gaps that are observed in failure occurrences in this status is considered the same
the previous articles are usually the conduction of preliminary occurrence however it will be calculated in
maintenance planning for a machine or a system the reduced time period.
mentioned in these articles. But as we know each system
consists of different components that these components Variables and parameters:
are different in terms of nature and function conditions.
For example, two different sections of a system have : The regular time unit production cost for product i
different failure rate, variable preventive and emergency ($/unit)
maintenance planning costs and different maintenance
and planning conduction period that should be considered : The overtime unit production cost for product i
in the modelling. In this article, all the different failure ($/unit)
conditions for a system is observed as an independent
system and the decision-making is carried out separately : The unit inventory cost to hold product i at the end of
for each one of them. each period for one period ($/unit)

II. MATHEMATICAL MODELLING : The backorder cost per unit of the product i in period
t ($/unit)
The aim of the introduced model in this article is
introducing an integrated model for the production : The labour cost in period t ($/man-hour)
planning and maintenance planning. Each of the
production machineries are exposed to different failures ℎ : Cost to hire one worker in period t ($/man-hour)

928
 Proceedings of the 2016 IEEE IEEM

: Cost to lay off one worker in period t ($/man-hour) : The binary variable; if the preventive maintenance
type j occurred in period t on machine m the value is
: Number of working days in period t 1 and otherwise it will be 0.
: The failure occurrence rate of type j on the machine
m (the number of events /day)
,

: The demand of product i in period t (units)

ℎ
: The labour time for product i (man-hour/unit)

: The machine time for product i (machine-hour/unit)
, ,
: The fraction of regular-time workforce available for
use in overtime in period t
1 (1)
: The maximum available workforce in period t , ,
(man-hour)
s.t
: The maximum machine capacity m in period t
(machine-hour) , , ; ∀ , (2)

: The preventive maintenance cost of type j on the ; ∀ (3)

machine m in time period t (currency)
; ∀ (4)
: The emergency maintenance cost of type j on the
machine m in time period t (currency) ; ∀ (5)

: The binary variable; if the failure type j has an

occurrence possibility for machine m the value is 1 ; ∀ (6)
and otherwise the value will be 0.

: The duration of the preventive maintenance of type j

on machine m in time period t (machine – hour)

: The duration of the emergency maintenance of type

j on machine m in time period t (machine – hour)

: The number of days wherein the failure rate of type j 1 ; ∀ , (7)
on machine m reduces to 0 after the preventive

maintenance.
, , , , , , 0 ; 1,2, … , ,
1,2, … ,
: The quantity of product i produced during regular
time in period t (units)

0,1 ;
1,2, … , , 1,2, … , , (8)
: The quantity of product i produced during overtime in 1,2, … ,
period t (units) In the expression 1 that relates to objective function
: The inventory of product i at the end of period t of this model, the total production and maintenance
(units) planning costs minimized. The second and third
expressions of the objective function depict the total
: The backorder level the product i in period t (units) workforce cost and total costs spent for increasing or
reducing the workforce level respectively. In the fourth
: The workforce level in period t (man- days) and fifth expressions, respectively the cost related to the
conduct or conduction lack of the preventive maintenance
: The number of workers hired in period t (man- days) planning is calculated. The constraint 2 states an
inventory balance equation of each product in each
: The number of workers fired in period t (man- days) planning period. In the constraint 3 balance equation of
workforce level is stated in each of the periods. In the

929
 Proceedings of the 2016 IEEE IEEM

constraint 4, the maximum workforce that can be adopted that the adoption of the proposed model in this study led
in each course is controlled. The constraints 5 and 6 are to a reduction in the production system costs. The
related to the maximum workforce capacity in each period introduced model here only involves an integration within
in the regular and overtime periods respectively. In the the environment of the production factory; this is while it
constraint 7, the production machineries capacity is can be extended on multi-level issues related to the supply
considered, so that the total periods when the machine is chain.
busy in the regular and also in overtime periods for
producing the related product, and besides the time spent REFERENCES
for emergency or preventive maintenance planning in a
period should not be more than the maximum capacity of [1] E. S. Buffa, and W. H. Taubert, “Evaluation of direct
existing machineries in that period. The constraint 8 computer search methods for the aggregate planning
shows the nature of the decision variables. problem,” Industrial Management Review, vol. 9, no. 1, pp.
19–28, 1967.
[2] G. Budai, R. Dekker, and R. P. Nicolai, “A review of
III. NUMERICAL EXAMPLE planning models for maintenance & production,”
Econometric Institute Research Papers (No. EI 2006-44),
For showing the applicability of the introduced 2006.
model, the information of Samen Pharmaceutical [3] G. Schmith, “Scheduling with limited machine
Company was used. This company produces 70 types of availability,” European Journal Of Operational Research,
different products in 5 independent production lines. In vol. 121, no. 1, pp. 1–15, 2000.
the current case study, the planning for 5 product families [4] L. Weinstein, and C. H. Chung, “Integration maintenance
among the mentioned products is considered including the and production decision in a hierarchical production
planning environment,” Computer and Operation Research,
bag serums with volumes of 50, 100, 500 cc, peritoneal
vol. 26, no. 10, pp. 1059–1074, 1999.
dialysis and bulky products, in a 6-months planning [5] E. H. Aghezaf, M. A. Jamali, and D. Ait-kadi, “An
horizon. The numerical example is solved by GAMS and integrated production and preventive maintenance model,”
the computational time is less than 5 Sec. European Journal of Operational Research, vol. 181, no. 2,
The aim of introducing an integrated model is pp. 679–685, 2007.
reduction of the total production system costs including [6] E. H. Aghezaf, and N. M. Najib, “Integrated production
the production cost and maintenance cost. In Table 1 the planning and preventive maintenance in deteriorating
results related to objective function values are expressed production systems,” Information Science, vol. 178, no. 17,
in both integrated and non-integrated states. pp. 3382–3392, 2008.
[7] S. Zhao, L. Wang, and Y. Zheng, “Integrating production
planning and maintenance: an iterative method,” Industrial
TABLE I Management and Data Systems, vol. 114, no. 2, pp. 162–
Objective function values in the integrated and non-integrated conditions 182, 2014.
Percent increase cost in relation to [8] M. B. Yildirim, and F.G. Nezami, “Integrated maintenance
Modelling type
integrated model and production planning with energy consumption and
Current condition 3% minimal repair,” International Journal of Advanced
Manufacturing Technology, vol. 74, no. 9-12, pp. 1419–
The entire failures are 1430, 2014.
implemented in emergency 12%
form
[9] M. Gehan, B. Castanier, and D. Lemoine, “Integration of
maintenance in the tactical production planning process
The entire failures are
implemented in preventive
under feasibility constraint,” in IFIP International
4% Conference on Advances in Production Management
form
Systems, pp. 467-474. Springer Berlin Heidelberg, 2014.

As it is specified in Table 1, the total system costs in

the non-integrated model are more than the integrated
model. As a result, the use of the integrated model causes
reduction of costs. Moreover, the introduced model leads
to reduction rate of 3% in the system costs compared to
the current condition.

IV. CONCLUSION

In this article, a mixed integer linear programming

model for the integration of aggregate production
planning and preventive maintenance planning was
introduced. In this model, we concluded that the
production planning and maintenance planning should be
considered in an integrated system. Likewise, the present
case study on Samen Pharmaceutical Company revealed

930