Filetype PDF | Posted on 05 Feb 2023 | 3 years ago
Authentication
digital resources
296x Filetype PDF File size 0.89 MB Source: fenix.tecnico.ulisboa.pt
The Optimization of Production Planning and Scheduling:
A Real Case Study in Ice-cream Industry
Mariana Pedroso Casal Ribeiro de Carvalho1
1 Master in Industrial Engineering and Management, Instituto Superior Técnico, Universidade Técnica de
Lisboa-UTL
Abstract
As other segments of the food industry, the ice-cream industry has its own features that influence the production
management of its processes. Amongst these we identify: changeover tasks, products shelf-life and perishability,
multiple deliveries during the planning horizon and RMs procurement and inventory control. These aspects have
been often left out when studying the production planning and scheduling within the batch food industries. Thus,
the problem seeks an optimal solution for the production planning and scheduling of a dairy food company, where
a methodology was proposed to address the main features of this industry based on the integration of two supply
contracts at the scheduling level. Two mixed integer linear programming (MILP) models are developed where the
RMs´ shelf-life aspect is integrated in the operation scheduling, extending them to explore different contractual
relationships between suppliers and the company, applying the RTN methodology. Both RTN mathematical
approaches are applied to the present case-study in order to evaluate which contract type is more suitable for the
present production process, in terms of scheduling, raw-material costs and final products’ quality. The obtained
results as well as the computational statistics are analysed.
Keywords: scheduling of production, perishable products, inventory control of RMs and supplier contracts.
1. Introduction cream industry is a food industry niche and the
The food industry has been growing throughout literature regarding this subject is very scarce.
the years. This growth has been driven by the Nevertheless, some research has been made
increase of new market competitors, as well as by an considering the problematic of the optimization of
the production planning and scheduling in food
increase on the consumers’ demand and industry. Entrup et al. (2005) have developed a
requirements. These factors combined have changed MILP formulation considering shelf-life restrictions
the trends in the industry and, production for the final products applied to the yoghurt industry.
optimization is nowadays a need. In this context the This aspect was accounted in the objective function,
production scheduling became an important activity which aims to maximize the margin contribution.
for companies, allowing them to determine when, Considering the segment of seafood products, Cai et
where and how a set of products should be produced al. (2008) proposed a formulation that takes into
considering the operational aspects. account the RMs perishability, considering three
Thus, the scheduling activity has an important type of decisions: i) the type of products to be
role for performance improvement while adjusting produced, ii) the processing time of resources to be
the resources consumption/production to demand allocated to each type of product, and finally, iii) the
through an accurate allocation. In the artisanal ice- sequence of products production . The authors stated
creams production, the major challenge is related to that this model can be applied to any system that has
the raw-materials’ quality aspects as the final limitations in terms of RM and uncertainty in the
products’ qualities/freshness is deeply dependent on delivery dates. Amorim et al. (2011) have also
these characteristics. Hence, it is important to presented a MILP formulation exploring two cases;
address this aspect at the scheduling level. i) a make-to-order and; ii) a hybrid make-to-order
RMs’ procurement and inventory control must /make-to-stock strategy.
also be considered as part of the different contractual Besides, another research stream has been
forms that exist between supplier and companies, explored considering the inventory models for
since in this type of industry the RMs price and perishable products such as the work developed by
availability suffers a high volatility which has a Goyal (1994) and Soman et al. (2004) based on the
direct impact in the way that RMs are purchased and application of the Economic Lot Size Problem
their costs. (ESLP). However, the ESLP has in its base
2. Related Literature assumptions that are unrealistic for the ice-cream
industry namely the constant demand rate assumed
A wide range of methodologies and techniques to in these problems which is not realistic for fresh food
deal with the production planning and scheduling is industries with seasonal products, as it is the ice-
available in the literature. However, the artisanal ice- cream industry.
1
Notwithstanding, the research on production
scheduling in the ice-cream industry with perishable Raw Materials Packaging Solidification
goods inventory control is still at its beginning. Reception
Moreover, since in the dairy industry the RMs’
procurement plays an important role, the relationship Intermediated
between companies and suppliers becomes a crucial Preparation Storage Frozen
aspect to be integrated in the planning and (Cooling) Storage
scheduling production mainly because of the
fluctuations on quality and price of RMs delivered. Blending Pasteurize/Ho
Thus two supply contract types are explored: the mogenize
Fixed Commitment Contract (FCC) and the Spot
Market Contract (SMC). The FCC is a long term Figure 1 – Generic Ice-cream/Sorbet Production Process.
contract where two aspects are defined a priori: i)
quantity supplied and ii) price of RMs. As for i), the 2.2. Methodology
quantities of RMs delivered in the period are fixed. The artisanal ice-cream industry has special
Consequently, in industries where demand is features, such as: changeover tasks between
uncertain, producers may take a significant risk as productions, multiple deliveries, shelf-life of
regards inventory management as they are unable to products, procurement and inventory control of
forecast the exact amount of RMs they will need. As RMs. All these features must be integrated in the
regards to ii), the price of RMs is negotiated by the scheduling process in order to make it, as close to
parties and remains the same during the contract. reality as possible. Moreover, due the importance of
Thus, the buyer does not benefit from the procurement and inventory control methods for this
fluctuations of RMs’ price which may occur in the work, it will be considered the integration of the two
spot market. However, the purchase price supply contracts at the scheduling level. To clarify
established is lower than the average spot market the integration of these aspects in the scheduling,
price. Other benefits can arise from a long term Figure 2, shows the methodology developed to be
commitment as Minner, S. (2003) points out. followed in the next chapters of this work.
The SMC is characterized both by allowing the
buyer to benefit from the fluctuations on RMs’ price
and by granting the flexibility of moving from one
supplier to another without any investment.
However, even though one can benefit from the
variations on price, in this type of contract the
company pays the spot market price.
Having this framework and exploring the
existent gap on the literature for addressing all these
features at the scheduling of production level, in the
present work are developed two mathematical
formulations for the artisanal ice-cream batch
multipurpose and multi-product process. The aim of
the models is to explore the characteristics of this
industry, accounting for RMs perishability and their
inventory control, changeover tasks and multiple
deliveries based on two different supply contracts, Figure 2 - Framework of the present case study.
simultaneously with the planning and process studystudy.
scheduling. The models consider the specific characteristics
of the present production process. The multiple
2.1. Description of Ice-cream Production Process deliveries and changeover tasks will be integrated at
The ice-cream industry has different types of ice- the same time that it is considered the perishability
creams production, which are often classified into of RMs. Taking into account the two types of
four categories according to their main ingredient: contracts, some features must be adapted to reflect
vanilla, cream, yoghurt and fruits/sorbets. However, the reality of each contract, which is the case of the
when considering the production process of these perishability of RMs. Thus, in the FCC, it must be
products they are mainly divided in two categories: integrated the control of RMs’ shelf-life in the
sorbets and other types of ice-creams production. scheduling of production to account with the reduced
Both categories differ in its production process until shelf-life of these products. In this sense, the RMs’
frozen task is reached. For confidentiality reasons a shelf-life control concept will be applied (Figure 2).
generic ice-cream production process is
characterized in Figure 1.
2
Selling price of ice-creams and sorbets;
Figure 3 - Schematic representation of RMs’ shelf-life. The production requirements and deliveries
dates along the planning horizon.
Determine:
The amount of each resource used;
The task-unit assignment and the batch size;
The optimal scheduling satisfying not only the
multi-deliveries along horizon, but also, the
The RMs’ shelf-life must be quantified from the demand at final horizon;
beginning of the planning horizon until the Raw-materials profile for the time horizon.
production instant. Considering instant t, the The final quantity of RMs discarded at the end
moment that a RM is required for production and, of the planning horizon by not respecting the
its shelf-life. If t ≤ its shelf-life is greater shelf-life restriction.
than the production date and the RM has the safety 3.1 Mathematical formulation for FCC
and quality properties to be processed. However, in In this section is presented the FCC model’s
periods that t > is verified, the the production
formulation. The problem can be defined as follows:
date t is greater than the RM shelf-life, it will not
fulfil the safety and quality patterns, going to the 3.1.1. Indexes
disposal. All the reaming quantity of RMs that have
already overcome its shelf-life and consequently, is d Deliveries
deteriorated, will be not used in production. k Tasks
Finally, for the SMC, this feature will not be r Resources
explicitly considered in the model since in this case t Time
the company will work in a JIT production strategy. θ Relative time to define the start of a task
3. Problem Statement
3.1.2. Sets
In this section with the objective of supporting the
production scheduling in an artisanal ice-cream Cr {rϵR: set of all material resources, such as:
production line of a multiproduct dairy batch plant, RMs, intermediate products and final
it will be developed and characterized both MILP products}
D {rϵR: set of all equipment and human
models based on the RTN approach. r
resources}
In most dairy industries, the quantity of final D {rϵD : the equipment resources used for
chg r
products to produce is known at the start of the the changeover tasks}
planning horizon. Some assumptions are considered K {k: Set of all tasks}
to develop the mathematical formulations, such as: proc
K {kϵK, rϵD : set of all processing tasks, k,
r
operating in an equipment resource, r}
All RMs used in the process are received at the Kchg {kϵK: set of all changeover tasks, k, to be
begging of the planning horizon. performed in an equipment resource, r}
The quantity of RMs received will fulfill the R {r: set of all resources}
weekly production. Rp {rϵCr: set of final products}
The RMs have different shelf-lives. Rrm {rϵCr: set of RMs}
The RMs are available for production until they Sr {rϵCr: set of material resources with
reach their shelf-life (t ≤ ). storage}
Τ {kϵK,rϵD : set of all tasks, k, that uses
hr r
The optimal scheduling of production can be operators teams as resources}
obtained by solving the following problem: 3.1.3. Parameters
Given: Spoilage cost of deteriorating RMs.
Process description by a RTN representation; Storage cost of deteriorating RMs.
The maximum amount of each type of resource H Planning horizon;
available; Selling price of final products;
Resource characteristics and capacities;
Acquisition cost of RMs;
Time horizon of planning;
min max
Q Q Minimum /Maximum amount of
Task and resources operating data; rd rd
RMs’ shelf-life; resource r Rp to deliver in instant d.
RM’s spoilage, purchase and storage costs; Rrdemand Demand of final resource r;
Quantities contracted of raw-materials; R0r Resource, rϵR, available initially;
3
Shelf-life of RMs; Binary Variables
ed ld N Binary variable that takes the value 1 if task k
T T The earliest/latest time to deliver, r kt
rd rd starts at time t, otherwise is 0;
R , in instant d.
p chg
min max N kt Binary variable that takes the value 1 if
Vkr Vkr Minimum/Maximum allowed
capacity of resource r to perform task changeover task k starts at time t, otherwise is 0;
k;
ℎ Processing time for task k/Processing Continuous Variables
Inv Raw-materials inventory level
time for changeover task k; rm
μchg Resource (processing unit) r ξkt Quantity of material undergoing task k at
krθ the beginning of instant t;
consumed at the start/end of the R Quantifies the amount of available resource
chg rt
changeover task k and at time r at the instant t;
relative θ; ∏ Total amount of final product, r, delivered
μ ν Resource produced/consumed of task rt
krθ / krθ at instant t.
k at the start/ end of time relative θ; IDisposal Quantity of RM r that has to be discarded
r
3.1.4. Variables by not respect the shelf-life restriction
In the model it is necessary to considered both Z Quantifies the profit
binary and continuous variables, as presented below.
3.1.5. Mathematical formulation for FCC Model
After being defined all the indexes, sets, parameters and variables, the mathematical formulation for the FCC
is developed, to be applied to the case study comprising the following constraints:
Objective Function
fp ( rm) Spoilage
= ∑ ∑(∏ × Price ) –[ ∑ ∑ R0 ×Price + ∑ (IDisposal × C )
r r r r
∈ ∈ ∈
∑ ∑( storage)
+r ∈ IDisposalr × C ] [1]
Subject to:
chg
τ
τk k
chg chg
Rrt = Rro|t=1 + Rrt−1|t>2 + ∑∑(μ Nkt−θ) + ∑ ∑(μ N )∀r ∈ DR,t ∈ H [2]
krθ krθ kt−θ
k θ=0 chg θ=0
kϵ k
τk
R =R0 + ∑ ∑(ν ξ ) ∀ r ∈ R , t = 1 [3]
rt r krθ kt−θ rm
k∈K θ=0
r
τk
Rrt = Rrt−1|t≥2 + ∑ ∑(νkrθξ ) ∀ r ∈ Rrm,t ≤ [4]
kt−θ
k∈K θ=0
r
τk τk
∑ ∑( ) ∑ ∑( ) ( )
R =R + ν ξ │ ≤ + ν ξ − −∏ ∀ r\R ,t ϵ H [5]
rt rt−1 krθ kt−θ krθ kt−θ rm
k∈K =Sto1 θ=0 k∈K \Sto θ=0
r k r 1k
IDisposal = R ∀ r ∈ R , t = −1 [6]
r rt rm
≤ ≤ ∀ ∈ , ∈ , ∈ [7]
∑(− ∏ )≥ ∀ ∈ , ∈ [8]
=1
≤ ∑ (−∏ )≤ ∀ ∈ , ∈ , ∈ [9]
=Ted
4
no reviews yet
Please Login to review.
