tudy of an innovative method based on complementarity between ARIZ,

lean management and discrete event simulation for solving warehousing

problems

Fatima Zahra Ben Moussa

a, , Roland De Guiob , Sbastien Duboisb , Ivana Rasovskac,

Rachid Benmoussa

a

a SyLPRO/National School of Applied Sciences of Marrakech, Avenue Abdelkrim Khattabi, 40000 Guliz, Marrakech, Morocco

b CSIP, ICube Laboratory, National Institute of Applied Sciences of Strasbourg, 24 Boulevard of Victory, 6700 Strasbourg, France

c ICUBE/University of Strasbourg, 4 Street Blaise Pascal, 67081 Strasbourg, France

ARTICLE INFO

Keywords:

Algorithm for Inventive Problem Solving

(ARIZ)

Lean management

Discrete event simulation

Witness

Design of experiments

Generalized system of contradictions

ABSTRACT

Warehousing challenges require not only effective management but also a new innovative strategy that can

significantly improve warehouse efficiency. This study presents an innovative method for solving warehouse

problems and achieving low-cost warehousing, based on complementarities between ARIZ (Algorithm for

Inventive Problem Solving), lean warehousing approach and discrete event simulation. In the proposed model,

the lean warehousing is first used to understand and analyze problems in order to propose solutions that improve

warehouse system efficiency. Then, a model of the improved system is developed and analyzed in a predictive

simulation software (Witness1 ), to evaluate its performance. The discrete event simulation allows for experi

ments on the system to be created and analyzed. Thus, an experimental design was developed to establish the

cause-and-effect relationships between the parameters of the system in order to formulate a generalized system

of contradictions. And finally, ARIZ is used to address related contradictions for searching for an innovative

solution, which must be subsequently implemented and evaluated in the discrete event simulation. The ap

plicability of the proposed method is demonstrated through a case study, involving an automotive company

specialized in the manufacturing of electronic modules that wants to reduce raw material warehousing costs.

Generally, the aim of the proposed method, which combined lean management, discrete event simulation and

the ARIZ algorithm, is to be able to propose solutions that go beyond the limits of continuous improvement and

optimization methods. Such an approach enables to propose a kind of continuum between traditional multi

objectives methods and inventive ones.

1. Introduction

In today's highly competitive industrial markets, warehouses in

creasingly play a significant role in supply chains. With the increasing

demand for good customer service, organizations are under pressure to

develop more efficient warehouses to satisfy customer requirements

and protect companies against variations in both productivity and de

mand rates (Drew, Mccallum, & Roggenhofer, 2008). As a result,

warehouses are now required to provide more flexibility, accessibility,

and efficient functionality to enhance organizational profits. The

warehouse plays a significant role in the supply chain, which delivers

the right quantity of the right product to the right customer, at the

optimal time, place, and price. A warehouse can be defined as a phy

sical storage system, from which products are received, transferred,

stored, picked up, sorted and accumulated, cross-docked, and shipped

(Cakmak, Gunay, Aybakan, & Tanyas, 2012).

Warehouse management, or warehousing, is the art of efficiently

operating a warehouse and is one of the key factors in supply chain

management (Lambert, Cooper, & Pagh, 1998). It is pertinent to men

tion that organizations face considerable challenges in managing

warehouses. Moreover, improving warehouse efficiency involves elim

inating any waste from a warehouse, streamlining its operations, and

improving the efficiency in every aspect of warehousing activities (Min,

2006). The successful application of continuous improvement

https://doi.org/10.1016/j.cie.2019.04.024

Received 24 March 2018; Received in revised form 22 March 2019; Accepted 15 April 2019

Corresponding author.

E-mail addresses: &..a@g(F.Z. Ben Moussa), r..o@insa-strasbourg.fr (R. De G..s@insa-strasbourg.fr (S. Dubois), i..a@unistra.fr (I. Rasovska).

1 https://www.lanner.com/en-gb/technology/witness-simulation-software.html.

Computers & Industrial Engineering 132 (2019) 124-140

Available online 16 April 2019

0360-8352/ 2019 Elsevier Ltd. All rights reserved.

Tmethodologies, especially "lean warehousing," would lead to sig

nificant improvements in a warehouse, such as the elimination of non

value-added processes, and an improvement in total lead-times and

customer satisfaction, with a zero-breakage rate for the warehouse.

In addition to continuous improvement strategies, another way to

improve warehouse efficiency is through the use of innovative strate

gies. Innovation can be recognized as a key success factor for increasing

competitiveness in a complex environment (Lambert et al., 1998). TRIZ

(The Theory of Inventive Problem Solving) is one of the most powerful

and widely accepted methods for enacting systematic innovation (Sheu

& Lee, 2011; Spreafico & Russo, 2016). TRIZ is a theory that has been

widely employed in multiple industries and domains to solve problems

and find inventive solutions. TRIZ provides a set of methods and tools

that are generally employed in preliminary technical product designs.

One of the main methods of TRIZ (the most complete one) is ARIZ

(Algorithm for Inventive Problem Solving). Historically, different ver

sions of ARIZ were proposed and tested through its use on many cases,

till ARIZ-85C, which could be recognized as the most mature meta

methods of TRIZ. ARIZ-85C clarifies the links between the different

tools and techniques of TRIZ, in order to apply them in a structured

manner. The TRIZ theory enables people to adopt a dialectical thinking

style, which guides them to understand problems as systems, con

ceptualize the ideal solution, and enhance the performance of the

system by solving contradictions, which are descriptions in TRIZ of

conflicts between problem objectives (Wang, Yeh, & Chu, 2016). TRIZ

was first proposed for solving technical problems related to product

engineering design. Subsequently, it was extended to resolve other

types of problems in other fields such as supply chains, service, edu

cation, and information processing.

Till now, no global approach to apply TRIZ methods in supply chain

has been proposed. One can recognize that previous research proposes

the integration of some TRIZ tools into various stages of DMAIC (Jiang

& Nguyen, 2015; Muruganantham, Krishnan, & Arun, 2014; Wang

et al., 2016; Xie & Li, 2009; Zhao & Zhao, 2013). These proposals were

primarily focused on product or process design, but not in management

of supply chain. Otherwise, there are limited studies in the existing

literature that deal with the application of TRIZ for problems in supply

chain operations (Chechurin & Borgianni, 2016). A previous study (Ben

Moussa, Rasovska, Dubois, De Guio, & Benmoussa, 2017) investigated

the application of TRIZ to supply chain problems, particularly to in

ventory management problems. They demonstrated that there is no

obvious or straightforward reason preventing TRIZ from being applied

to all categories of supply chain problems, including warehousing

problems. Also, the authors in Ben Moussa et al. (2017) highlight the

limitation of using the most mature meta-methods of TRIZ (ARIZ-85C)

in solving unconventional generalized contradictions. Within this con

text, the aim of this paper is to study the use of the ARIZ algorithm in

solving problems formulated by generalized contradictions. However,

using ARIZ-85C for solving problems is sometimes difficult because

TRIZ theory lacks highly specific tools and methods to understand

complex problems and formulate a related system of contradictions. For

this reason, in this paper, the authors study the synergetic application

and complementarity of lean warehousing, discrete event simulation

and ARIZ 85C, and propose a framework based on these three methods.

In this method, lean warehousing and discrete event simulation are

used to provide the necessary data to formulate relevant systems of

contradictions that prevent the fulfillment of warehouse efficiency ob

jectives, whereas ARIZ-85C is adopted to address the formulated con

tradictions. Unlike the objectives of lean warehousing and discrete

event simulation, which are to improve warehouse efficiency and

search for best compromises between the problem constraints, TRIZ

allows to move beyond compromises by overcoming contradictions.

Then, the complementarity of lean warehousing, discrete event simu

lation and ARIZ-85C allows better results and performances to be ob

tained than solving the corresponding problems using only lean ware

housing or simulation software. Thus, this study contributes to existing

works by using the lean warehousing methodology in conjunction with

the inventive algorithm ARIZ-85C and highlights the application of

ARIZ-85C to solve problems formulated by generalized technical and

physical contradictions, as there is no existing literature that deals with

this subject.

The remainder of this paper is organized as follows. Section 2 in

troduces the materials and methods, including lean warehousing, ARIZ

algorithm, generalized system of contradictions and a literature review

of papers relating to the application of TRIZ tools with the lean-ware

housing and DMAIC. Section 3 presents the proposed innovative

method based on the complementarity of ARIZ, simulation and lean

warehousing. Section 4 presents the result of applying the proposed

method to a warehousing problem through a case study involving an

automotive supplier producing electronic modules. A discussion is

provided in Section 5, and conclusions and future research are dis

cussed in Section 6.

2. Materials and methods

2.1. Lean warehousing

Warehouses comprise a substantial component of logistic opera

tions, and they play a critical role in matching product demand with

supply across different echelons in a supply chain. Today, warehouses

represent centers for value-addition rather than centers for storage

(Johnson & McGinnis, 2010). Moreover, many warehouses continue to

suffer from significant inefficiencies, as forklift operators waste time

and resources hunting and digging because they lack adequate in

formation on the locations of items and optimal routes for storage, re

plenishment, and retrieval actions (Hackman, Frazelle, Griffin, Griffin,

& Vlasta, 2001). Warehousing activities/operations include receiving,

unpacking/sorting, storing, picking up, sorting, packing, and shipping

(Aminoff, Kettunen, & PajunenMuhonen, 2002). Furthermore, Bottani,

Cecconi, Vignali, and Montanari (2012) have shown that order picking

is the most labor-intensive and costly activity of most warehouses, with

approximately 55% of the total warehouse operating expenses relating

to order picking operations. To improve a warehouse's efficiency and

optimize its operations, many methods have been presented in the lit

erature, such as routing methods to determine the sequences and routes

of picking, storage assignment methods to assign items to storage lo

cations based on certain rules, order batching algorithms to group two

or more customer orders into one picking order, warehouse layout

design to minimize total travel distances, and methods of the lean

management philosophy to eliminate waste from warehousing activ

ities and non-value added processes, and improve the total lead-times to

achieve low-cost warehousing.

The lean management philosophy is the most widely known ap

proach to industrial continuous improvement, and comprises an op

erations management approach that refers to non-value adding activ

ities as waste. The key philosophy of lean management is that the

elimination of non-value adding activities, variability, and inflexibility

is imperative in order to deliver value to customers with the right time,

quantity, and quality and at the minimum cost (Drew et al., 2008). The

implementation of lean management leads to the removal of eight types

of waste: transport, inventory, motion, waiting, over-processing, over

production, defects, and talent (Kadarova & Demecko, 2016).

The subject of lean warehousing deals with the application of lean

concepts and practices in warehousing operations to improve the

warehouse efficiency. Furthermore, the benefits associated with lean

warehousing include an improved quality of operations, improved

profitability, and a leading edge in the market competition (Mustafa,

2015). To this end, lean warehousing includes numerous tools and

techniques, for instance VSM (value stream mapping), Mudas, root

cause analysis, and activity diagrams. The application and choice of

lean tools depends on the user's experience and knowledge. Further

more, when applying lean management or lean warehousing to

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

125complex processes, the use of a structured method for problem solving,

such as DMAIC (Define, Measure, Analyze, Improve, Control), is

strongly recommended. DMAIC is a structured problem-solving method

and roadmap that can be used for any project, or with the goal of de

livering supply chain improvements (Bottani et al., 2012). The term

DMAIC stands for the five main steps in the process: identifying the

problem, measuring the current status, analyzing where the problem is,

improving the performance, and controlling the new process.

2.2. Algorithm for inventive problem solving (ARIZ)

TRIZ, which is the Russian acronym for the Theory of Inventive

Problem Solving, was developed by Genrich Altshuller starting in 1940.

Unlike problem solving methods that accept randomness in the in

novation process, TRIZ is a knowledge-based systematic methodology,

which provides a logical approach to develop creativity for innovation

and inventive problem solving, while rejecting compromises between

objective restrictions and specific situation limits (Ilevbare, Probert, &

Phaal, 2013). TRIZ is based on the three fundamental concepts of re

sources, ideality, and contradictions. Ideality is one of the most pow

erful concepts of TRIZ. The meaning of this ideality is to maximize the

constraints of the considered problem in order to find the more robust

solutions. According to ideality, the ideal final result consists of solving

the problem without introducing new resources, and searching for

configurations of the system where the desired result is achieved by

itself. An ideal system may not be possible to achieve, but knowledge of

the ideal system helps in improving an existing system, by maintaining

the features of the ideal system as a goal to aim for. Contradictions can

be either technical or physical, and one hypothesis of TRIZ is that such

contradictions exist as soon as a problem exists without a known so

lution. Technical and Physical Contradictions are just two different

models to represent the same problem, but they always exist together.

In addition to the key concepts of TRIZ, Altshuller developed a set of

methods, tools, and a knowledge base for systematically generating

new ideas and solutions for problems that arise during the evolution of

technical systems. Subsequently, further studies based on TRIZ have

been conducted in order to extend TRIZ approaches to solve other kinds

of problems, not just those related to the evolution of technical systems

(Khomenko, De Guio, & Cavallucci, 2009).

ARIZ is the Russian acronym for the "Algorithm for Inventive

Problem Solving," developed by the TRIZ creator Genrich Altshuller.

ARIZ is a process that links the tools and techniques of TRIZ, in order to

apply them in a structured manner, to evolve from a complex, fuzzy,

problem to a point where it can be solved. Throughout its evolution,

several methods and tools were defined in TRIZ. To avoid considering

TRIZ as a tool box but to well structure the use of these methods and

tools in a systemic way, it was necessary to build a meta-method to

guide the designers to combine them in a relevant and robust way. ARIZ

has been developed through several versions, and the last one proposed

by Altshuller is ARIZ-85C (Fiorineschi, Frillici, & Rissone, 2015). The

framework of ARIZ-85C, illustrated in Fig. 1, consists of nine algo

rithmic parts (Altshuller, 1985). These parts can be selectively em

ployed according to the practical situation of solving a problem. Parts 1

to 4 are dedicated to analyzing the problem situation by converting the

initial problem into a formulated description, analyzing the problem

model by identifying the existing resources for solving the problem,

defining the ideal final result and the physical contradictions, and

mobilizing and using substance-field resources by increasing the

availability of resources by modifying the previously defined existing

resources, respectively. Scientific effects and a knowledge base are

applied in part 5. If the problem remains unsolved at the end of part 5,

we move to part 6 to change or substitute the problem. Otherwise, we

move to the last parts (parts from 7 to 9), which are dedicated to

analyzing the method of resolving the problem, applying the obtained

solutions, and analyzing the entire problem-solving process, respec

tively.

It is important to note that ARIZ-85C only addresses the re

formulation and resolution of the problem. Indeed its last version,

ARIZ-85C, the method has been developed to clearly state a problem

and to use cognitive tools to find a solution to this problem. But, con

trary to the previous versions, it lacks specific tools and methods to

analyze the problematic situation and to identify the prior problem to

consider. Contradictions are the pillar of ARIZ, and solving an inventive

problem means solving the related contradictions. Two steps of con

tradiction formulation are proposed. A Technical Contradiction is the

starting point of the whole problem-solving process using ARIZ.

Sometimes, the Technical Contradiction within a problem is clearly

evident, and at other times it seems that a problem does not contain any

Technical Contradiction, because it is hidden within the problem con

ditions (Davide & Montecchi Tiziano, 2012). This point reinforces the

necessity to well analyze the situation at the beginning of problem

solving process, to be sure to identify the good contradictions to con

sider. The second level of contradiction formulation is the Physical one,

stating the core of problem and explaining why the problem exists, and

why the system presents contradictory requirements.

2.3. Generalized system of contradictions

2.3.1. Classical TRIZ contradictions

One of the main ideas of TRIZ is to guide the evolution of systems by

identifying the contradictions that have to be overcome (Dubois,

Rasovska, & Guio, 2008). Contradictions can be either technical or

physical. Technical Contradictions (TC) appear when there are con

flicting requirements regarding two Evaluation Parameters (EP) of a

technical system. Physical Contradictions (PhC) appear when the same

Action Parameter (AP) should exhibit different properties at the same

time (Fresner, Jantschgi, Birkel, Brnthaler, & Krenn, 2010). Based on

the two contradiction models (technical and physical), a system of

contradictions was introduced by Khomenko et al. (2009) to represent

the causal relation between the AP and EP, as illustrated in Fig. 2. This

system of contradictions is based on the existence of a contradiction of

the parameter and a contradiction of the system, which justifies the

need for two different states of the parameter (Dubois, Rasovska, &

Guio, 2009).

Example. In an inventory management system managed by a Kanban

card system, a conflict between the stock size (capacity) and the

customer level service rate appears when the need for some products

increases (see Fig. 3). Indeed, the problem appears because there is a

lack of space to store an extra quantity in the stock. Moreover, if the

service concerned by this increase in demand does not receive the

necessary quantity, there will be a stock-out. To solve this problem, the

increased requested quantity must arrive to the stock without

increasing the storage space and without creating stock-out. To reach

this desired result, the reorder quantity of the Kanban card should be

big and small at same time, which is contradictory. This contradiction

where one parameter has to take two different values at the same time

is the so called physical contradiction in classical TRIZ.

2.3.2. Generalized system of contradictions model

The classical TRIZ contradiction is limited by the number of eva

luation parameters, which does not exceed two parameters. Indeed,

when dealing with real and complex problems, this model cannot be

used to represent the system of contradictions, owing to the problem

multidimensionality. In order to represent the system of contradictions

for a multidimensional problem, some improvements of the classical

contradictions models were proposed in the literature, such as the

Multiple-to-Multiple Parameter Contradictions (Sheu & Chen, 2011)

and the Generalized System of Contradictions (Dubois et al., 2009;

Dubois, De Guio, & Rasovska, 2011). Indeed, the Generalized System of

Contradictions is a generalization of the classical contradictions of

TRIZ, which links two Generalized Technical Contradictions (GTC) and

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

126a Generalized Physical Contradiction (GPC) (see Fig. 4). In the model of

GPC, the single Action Parameter and its values (defined in the classical

PC) are replaced by a set of Action Parameters and two concepts (two

combinations of states of the related APs). The generalization of the

Technical Contradiction is built on two concepts, two sets of Evaluation

Parameters, which are satisfied or not. The model of GTC does not rely

Fig. 1. Framework of ARIZ (according to (Altshuller, 1985)).

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

127only on two EPs, but generalized the definition, by emphasizing the

impossibility to satisfy simultaneously two set of EPs. Thus, the Gen

eralized System of Contradictions represents the generalization of the

classical TRIZ system of contradictions, where two concepts based on a

set of action parameters satisfy two sets of evaluation parameters. The

desired result is then the simultaneous satisfaction of the two sets of

evaluation parameters.

Example. Let's take the same inventory management system presented

in the previous section. A solution that focuses on increasing the

number of Kanban cards has been proposed. But, when solving the

contradiction presented in Fig. 3 by applying this solution, another

evaluation parameter is deteriorated (the transportation cost). Thus,

when the stock size and the stock-out meet the problem requirements,

the transportation cost does not meet the requirements. To solve this

system of contradictions (presented in Fig. 5), the stock size, the stock

out and the transportation cost should meet the problem requirements.

2.4. Lean-DMAIC and ARIZ

In the existing literature, one can find proposals of integration of

TRIZ with methods such as lean warehousing and DMAIC. In

Muruganantham, Krishnan, and Arun (2013) and Muruganantham et al.

(2014), the authors propose a synergetic approach of lean methods with

TRIZ to solve problems in a more effective way and to obtain better

results. The proposed approach was applied to reduce costs for manu

facturing components and to improve the productivity of industrial

processes. In this approach, lean methods are used to identify waste and

to identify and alleviate problems; TRIZ is employed to find the optimal

approach to reduce waste and to solve problems by resolving Technical

Contradictions through the use of one of the most widespread TRIZ tool,

the 40 inventive principles and its related matrix. These principles have

been extracted from the analysis of technical patents, and the matrix is

a statistical organization of the principles to propose the most used ones

for a given Technical Contradiction. The application of these principles

to non-technical problem, and more over the relevancy of the statistical

proposal, begs question. In Jiang and Nguyen (2015), the authors in

tegrate the TRIZ methodology with the "improve" phase of the lean

DMAIC approach, as a quality improvement strategy. Likewise, in the

proposed method, the authors employ only the tools dedicated to sol

ving Technical Contradictions, the contradiction matrix and the in

ventive principles. In Wang et al. (2016), the authors integrate TRIZ

methodology with the "analyze" phase of the DMAIC approach to re

volutionize the manner in which companies develop new products. In

the "analyze" phase, three basic TRIZ tools are used: (1) the engineering

parameters to describe technical conflicts and inventive principles to

solve technical contradictions, (2) a knowledge database of scientific

effects (physical, geometrical, and chemical) to solve problems, and (3)

the substance-field model for modeling technological problems. In

Tatjana Sibalija and Vidosav (2009); Xie and Li (2009); Zhao and Zhao

(2013), the authors propose incorporating TRIZ into various stages of

DMAIC to integrate the ideas of TRIZ with DMAIC. The objective of this

integration is to increase the effectiveness of DMAIC deployments.

From the existing literature, one can recognize that the proposed

methods in the cited papers share a common point, which is the in

tegration of some TRIZ tools into the DMAIC approach to improve the

overall potential of DMAIC for existing systems. In this article, the

authors aim at proposing a global approach, relying on the com

plementarities of ARIZ, lean warehousing and simulation tools, to solve

warehousing problems. This approach is proposed in the next section.

This proposal starts with a traditional resolution of problems with lean

warehousing, an optimization of the proposed solutions with simula

tion, and finally, if not satisfying, considering these optimized solutions

as the entry point to apply ARIZ.

2.5. Research methodology

Within the context of systematic innovation in the supply chain

areas, this paper aims at studying the use of the ARIZ-85C algorithm to

solve complex problems, formulated as a Generalized System of

Contradictions. This original proposal will also be illustrated through

an industrial case study, to validate the applicability of ARIZ-85C to

solve Generalized System of Contradictions. The chosen case study is

related to the warehouse area, thus distinguishing itself from the ma

jority of other studies (Ben Moussa et al., 2017).

3. Outline of the proposed methodology

Based on the lean warehousing methodology, following the DMAIC

approach, the discrete event simulation and the algorithm for inventive

problem solving (ARIZ) given above, the proposed model is illustrated

Fig. 2. Classical TRIZ system of contradictions.

Fig. 3. Example of a classical TRIZ system of contradictions.

Fig. 4. Generalized system of contradictions.

Fig. 5. Example of a generalized system of contradictions.

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

128in Fig. 6. In this proposed model, lean warehousing is used to improve

the system, discrete event simulation software is used to optimize the

improved system and bring the problem to a point where the for

mulation of contradictions preventing the attainment of the problem

objectives can be achieved, and ARIZ is used to solve these contra

dictions. One of the main questions related to this proposed metho

dology is that ARIZ has been defined to treat classical TRIZ contra

dictions. But, as will be illustrated, there is no difficulty to apply this

generic algorithm with the Generalized System of Contradictions.

Step 1: Apply the methodology "lean warehousing- DMAIC"

The main purpose of step 1 is to improve the warehouse and its ef

ficiency by eliminating waste from the material handling activities and

non-value-added processes. Lean warehousing is a useful process im

provement methodology, which is applied to warehouses and described

as a systematic methodology for eliminating waste and reducing the

complexity of a process. It is employed in the first step of the proposed

methodology to optimize the performance of a warehouse. In general, the

DMAIC cycle is employed to solve problems via a lean approach by

following the different phases of DMAIC in order to frame and structure

any lean implementation project. However, the choice of lean manage

ment tools to apply always depends on the problem to be solved. The lean

management toolbox contains a wide range of tools and methods, to re

spond to any situation in a lean management project. The choice of ap

propriate tools and methods depends on the project requirements.

Moreover, the control step of DMAIC is not applied in the proposed

methodology, as the objective of the proposed "ARIZ based lean ware

housing and computer aided simulation" method is to look for more

opportunities for problem improvement. Furthermore, the evaluation of

the proposed concept of solutions is performed with the aid of simulation.

Step 2: Evaluate the solutions using discrete event simulators

The main objective of this step is to implement the concept solutions

and provide the decision variables for the simulation program to con

firm that there is a measurable improvement. To do so, a model of the

future warehouse system of the studied problem is implemented on a

discrete event simulator (Witness 14 is used here). The simulation

program will run the model and provide results for the objective

Fig. 6. ARIZ based lean warehousing and computer aided simulation.

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

129function, to evaluate the future system efficiency. The chosen simula

tion software (Witness 14) provides tools for random experiments,

called 'experimenter' to find the best solutions (the Pareto front) for the

problem in the solution space.

Step 3: Use the design of experiments method

The Design of Experiments (DoE) is a systematic method for de

termining the relation between the input and output parameters of a

process. This information is required to manage process inputs in order

to optimize the output.

In this step, we propose the realization of two designs of experiments:

The design of experiments of the initial situation (DoE 1) and design of

experiments of the improved situation (DoE 2). Each DoE is characterized

by (1) a set of action parameters X = (X1, ..., Xk), (2) a set of evaluation

parameters Y = (Y1, ..., Yn), and (3) a set of experiments E = (E1, ..., Ep).

Each experiment Ei is characterized by a set of values (Vi1, ..., Vik) at

tributed to the set of action parameters, and by a set of values

(Zi1, ..., Zin) taken by the evaluation parameters, as listed in Table 1.

The realization of DoE1 and DoE2 will help to clearly visualize the

difference between the two situations, and compare their performances.

Step 4: Extract and formulate the generalized technical contra

dictions of the problem

The extraction and the formulation of a generalized system of

contradictions is one of the most important elements in the proposed

methodology. The generalized system of contradictions represents the

result of the analysis performed in the previous steps, and is the entry

point for the next step, which deals with the application of ARIZ.

After realizing two designs of experiments, one for the initial si

tuation (DoE1) and the other for the improved situation (DoE2), the

next step is to transform each DoE response matrix (Zij) into a binary

form, in which all the variables are restricted to binary values. Thus, Zij

takes the value of 1 if the evaluation parameter Yj is satisfied by the

experiment Ei and Zij takes the value of 0 otherwise. If no experiment Ei

enables the satisfaction of all the evaluation parameters Yj, this means

that no solution can be found in Table 1 for the performed DoE. Con

sequently, a Pareto set with at least two points exists for this problem,

and a classical TRIZ system of contradictions, or a generalized system of

contradictions, can be formulated. First, we start by creating the rela

tions of dominance between the different experiments Ei, in order to

find the non-dominated solutions from DoE1 and DoE2 (Pareto front).

Following the well-known concept of Pareto dominance, one solution

(S1) is better than another (S2), or S1 dominates S2, if the set of values

taken by the evaluation parameters for S1 is better than the set of values

taken by the evaluation parameters for S2. If the Pareto front is com

posed of solutions from DoE1 and DoE2, then a generalized system of

contradictions exists. If not, only a classical TRIZ system of contra

dictions exists. Finding the best generalized technical contradictions

involves finding the existing conflicts between the evaluation para

meters from the non-dominated solutions. To complete the description

of the system of contradictions, it is necessary to find the generalized

physical contradiction (GPC) that leads to the generalized technical

contradictions (GTCs). The GPC is composed of two concepts, based on

a set of action parameters that characterize the experiments of the non

dominated solutions (see Fig. 4).

Step 5: Apply ARIZ to address the contradictions

The main objective of this step is to propose concepts of solutions

for the studied problem, through the resolution of the physical con

tradiction or the generalized physical contradiction identified in Step 4.

The physical contradictions are conflicts among the design variables

(action parameters of the DoE) explaining (causing) the Pareto frontier

(effect). This model is different from the classical lean cause-effect

models, in that it analyzes conflicts of causes resulting in conflicts of

effects. In the previous step, we proposed formulating the system of

contradictions. This information allows the transition from an initial

problem situation to a clearly formulated and simplified description of

the problem. The identification of available resources may be useful for

solving the problem. The Ideal Final Result (IFR) and the Physical

Contradiction (PhC) that prevents the achievement of the IFR should be

formulated. The existing TRIZ principles should be used to solve the

formulated physical contradiction.

4. Results from the application of ARIZ based lean warehousing

and computer aided simulation method

In this section, we present a case study to demonstrate the applic

ability of the proposed method (ARIZ based lean warehousing and

computer aided simulation). The case study focuses on the reduction of

the raw material warehousing costs of an automotive company, spe

cializing in the manufacturing of electronic modules, by reducing the

number of operators working in the warehouse. The plant is composed

of two production areas called the SMT area and power area. The fac

tory operates for 24 h a day, six days a week. To ensure the production

continuity of the two areas, 48 operators (called feeders) work every

day to supply the various production lines with the necessary raw

materials to manufacture the electronic modules. The 48 feeders are

divided into three shifts (16 operators per shift, working eight hours a

day). Depending on the nature of their tasks, the feeder operators are

divided into four teams:

- team for reception and storage of raw material (two operators per

shift)

- team for reception of the raw material used in the SMT production

area (three operators per shift)

- team for raw material preparation for the SMT production area

(seven per shift)

- team for raw material preparation team for the power production

area (four operators per shift)

The storage warehouse for raw materials is composed of three main

areas: a reception area for raw material, a storage area for raw material

intended for the "SMT" production area, and a storage area for raw

material intended for the "power" production area. The preparation of

raw material is carried out in trolleys.

The company is implementing a downsizing policy. Thus, the supply

chain manager wishes to reduce the number of operators working in the

Table 1

Design of experiments matrix.

X1

X2

...

Xk-1

Xk

Y1

Y2

...

Yn-1

Yn

E1

V1,1

V1,2

V1,k-1

V1,k

Z1,1

Z1,2

Z1,n-1

Z1,n

E2

V2,1

V2,2

V2, k-1

V2, k

Z 2,1

Z 2,2

Z 2, n-1

Z 2, n

E3

V3,1

V3,2

V3, k-1

V3, k

Z 3,1

Z 3,2

Z 3, n-1

Z 3, n

Ep-2

Vp-2,1

Vp-2,2

Vp-2, k-1

Vp-2, k

Z p-2,1

Z p-2,2

Z p-2, n-1

Z p-2, n

Ep-1

Vp-1,1

Vp-1,2

Vp-1, k-1

Vp-1, k

Z p-1,1

Z p-1,2

Z p-1, n-1

Z p-1, n

Ep

Vp,1

Vp,2

Vp, k-1

Vp, k

Z p,1

Z p,2

Z p, n-1

Z p, n

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

130warehouse, as he considers that the work necessary for reception and

preparation does not require such a large number of operators. On the

contrary, the team leaders of feeder operators ask to increase their

numbers, because they fail to complete their tasks in time without the

intervention of external resources.

According to the ARIZ based lean warehousing and computer

aided simulation model, the implementation results are presented

below.

4.1. Step 1: Apply the methodology "lean warehousing- DMAIC"

4.1.1. Step 1.1: Define the problem statement (define)

In this step, the five W's and one H "5w1h" method (Table 2) is first

applied to define the problem. Then, in order to define the process/

needs of the process customer, the SIPOC "Supplier Input Process

Output Customer" diagram (Fig. 7) is employed.

- 5w1h method:

- SIPOC diagram:

4.1.2. Step 1.2: Measure the current state of the problem (measure)

The main objectives of this section are documenting the current

situation of the activities of the operators working in the raw material

warehouse and identifying their non-value-added activities (Mudas).

- The Business Process Model (BPM) diagram

The reception of the raw material is carried out in the reception area

by the reception and storage team, and then the raw material is con

trolled by the quality control team. Then, raw material for power is

addressed in the storage area for raw material intended for the "power"

production area (PA), and the raw material for SMT is addressed in its

temporary location. Next, operators of the reception team of the RM

used in the SMT PA transmit the boxes of RM, unpack them, check the

references (check the labels of each piece or set of RM to ensure that the

arrived product corresponds to the one noted on the box), and then

store them in the racks.

In the storage warehouse intended for the SMT PA, there are 19

trolleys to fill. The trolleys are localized in the trolley zone, and the raw

material is stocked in several racks. The trolleys are composed of a large

number of empty slots (up to 280 places), which are characterized by

the reference of the RM to be put in place. To fill the trolleys, operators

make round trips from the trolley zone to the storage racks to pick up

the necessary RM. In a round trip, an operator picks up only one re

ference. However, in the storage warehouse intended for the power PA,

the trolleys are mobile, and operators push them during their picking

up operations.

All the operations performed by operators are managed by a system

with barcode labels (with reader guns), linked with the ERP of the

company. Stock status is known throughout the system. Thus, any

product, its quantities, and its locations are recorded on the system.

By using the BPM diagram, we describe the activities of operators

working in the warehouse described above in detail (Fig. 8).

- The identification of waste/Mudas:

The identification of Mudas (non-value-added activities) is the pri

mary goal of this step. Mudas can be classified into several categories,

which include transportation, waiting time, unnecessary movement,

inventory, over-processing, and defects. A field analysis for the opera

tors' activities was carried out to identify the Mudas. Then, a time

measurement was performed to calculate the time spent in each Muda.

The following table classifies the identified Mudas by category.

Table 2

5w1h to define the problem.

what What is the problem?

- Reduce the number of operators (feeders) working in the warehouse.

who Who has the problem?

- The supply chain manager.

where Where is it happening?

- In the storage warehouse for raw material.

when When is it happening?

- 24 h a day, seven days a week, when the factory is in production

activity.

how How can this problem be overcome?

- Use the lean management approach to solve the problem and improve

the main process.

why Why solve this problem? What are the quantified issues?

- To eliminate waste from the material handling activities.

- To organize the warehouse's operations.

- To achieve low-cost warehousing.

Suppliers of the raw

material

Raw material

Delivery form

Feeding trolleys

filled by the raw

material

Production lines

Reception and

control of the

raw material

Storage

of the raw

material

Preparation of

the raw

material in the

feeding

trolleys (order

picking)

bring the

feeding

trolleys to the

production

site

Inputs

Suppliers

Process

Outputs

Customers

Fig. 7. The SIPOC diagram.

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

131From Table 3, the most expensive non-value-added activity is the

"unnecessary travel/round trips" in the storage area intended for the

SMT during the picking-up operations. This waste costs 37 h a day,

divided between the 21 feeders (seven operators per shift) for the SMT

production area. The next step will focus on the elimination of the se

lected Muda.

Fig. 8. The BPM diagram describing the activities of operators.

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

1324.1.3. Step 1.3: Analyze the root causes of the problem (analyze)

In this step, a cause and effect analyses is performed using the

Ishikawa diagram to determine the principal root causes of the most

expensive Muda, "unnecessary travels/round trip in the storage area in

tended for the SMT production area," identified in the previous step

(Fig. 9). The Ishikawa diagram identifies and maps out many possible

causes for a problem, related to several categories, from a brain

storming session.

4.1.4. Step 1.4: Identify possible improvement actions to eliminate the root

causes of the problem (improve)

After identifying the root causes of the problem, the problem-sol

ving step to propose improvement solutions follows. To generate ideas

for improvement or concept solutions, tools such as brainstorming,

creativity techniques, lean management tools, and specific warehouse

optimization methods (see Section 2.1) can be employed. Indeed, there

exist four methods for order picking optimization to reduce the travel

distance of an order-picker: (1) storage location assignment, (2) ware

house layout design, (3) order batching algorithms and (4) pick-routing

methods. To solve the present problematic situation, the authors focus

on the pick routing and the layout design methods. Thus, it is proposed

to change the order picking strategy from a round trip strategy to a

traversal strategy, and make some changes to the warehouse layout to

minimize the total travel distance. In the initial situation, an operator

fills the trolleys of the raw material one by one, by moving back and

forth from the location of the trolleys ("trolley zone") to the storage

racks where the product to pick up is stored. In this situation, the op

erators adopt a random picking sequence. Indeed, the order in which

the products are picked up has no influence on the total workload of the

operators. In the improved situation, it is proposed that operators

Table 3

Identification of Mudas.

Mudas

Processes

Feeders for the SMT production area

Feeders for the Power production area

Transportation

Waiting time

Unnecessary movement

In the storage area for SMT, operators do not move the trolleys during

the picking operations. They make several round trips from the trolley

area to the storage racks to collect the raw material needed to fill each

trolley. The number of round trips is equal to the number of product

references to put on the trolley.

35 h a day

In the storage area for power, operators move a lot between the

storage racks to prepare the RM. operators can go around the store

every time to find the component to be picked up.

13 h a day

Inventory

Over-processing, unnecessary

treatments, additional

control

A weekly inventory is done for the SMT storage area. 27 operators

participate in this inventory. Each operator spends an average of

45 min on the inventory.

(27 op * 45 min) = 20.25 h (per week)

An inventory is done three times a week for the storage area for

power. 15 operators participate in this inventory. Each operator

spends an average of 25 min in the inventory.

(15 op * 25 min) = 18.75 h (per week)

Defect

The operators sometimes forget to scan the barcode of a component to

declare its new location during the operations of picking up. This

generates a bad image of the stock on the information system.

The operators sometimes forget to scan the barcode of a component to

declare its new location during the operations of picking up. This

generates a bad image of the stock on the information system.

Unnecessary

travels in the

storage area

intended for the

"SMT"

Method

Workers

Materials

Environment

Part numbers are pasted on the trolley

A type of RM is always stored in different locations.

The operator does not move the trolleys

Operator does not have pick list

Random storage of the RM

The space between the storage racks does

not allow more than two trolleys to pass at

the same time

existence of a

great variety of

types of the

raw material

several

operators work

at the same

time on filling

trolleys with

raw material

Fig. 9. Ishikawa diagram.

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

133perform a warehouse tour between the storage racks by pushing the

trolley to fill and following the shortest path through the storage racks

(see Fig. 10). Each trolley must be filled in one journey between the

racks, while the starting and ending point of this journey is the trolley

area. To achieve this, operators must have a pick-up list that indicates

the order picking to follow for each trolley. As the raw material is

randomly stored in the warehouse, the picking sequence will be defined

based on the locations of products in the storage racks according to the

information system. This strategy leads to a route in which the racks,

that are to be visited, are totally traversed. In this improved situation,

the trolley becomes mobile, and the picking sequence becomes pre

defined. As part of the same improvement action, we make minor

changes to the warehouse layout.

4.2. Step 2: Evaluate the solutions with simulators

This step presents the results from the implementation of the pro

posed solution. The model of the future warehouse system for the pro

blem has been implemented on a discrete event simulator (Witness 14),

as shown in Fig. 10. In order to simulate the system, the identification of

the different parameters of the problem must be performed at this stage,

based on the problem objectives. Thus, the evaluation parameters of the

problem and the action parameters that influence these are:

PE1: delay to fill the 19 trolleys must be less than or equal to five

hours, to create a 3-hour availability to operators, in order to assign

them new tasks to perform.

PE2: total load for operators be less than or equal to 20 h, at a rate of

one hour per trolley.

PE3: medium capacity reserve (which represents the available time of

an operator after completing his tasks of filling the trolleys with raw

material) must be greater than or equal to three hours, to assign to

operators new tasks to perform.

PE4: total waiting time must be equal to zero minutes.

PE5: total travel time of operators must be less than two hours, at a

rate of two full warehouse tours for each trolley

PA1: picking sequence,

PA2: number of available operators,

PA3: type of trolleys.

The objective of simulating the new system is to compute its per

formance, in order to examine whether there are measurable im

provements or not compared with the initial system. Furthermore, this

simulation will allow the optimal solution (the Pareto optimum) for the

problem to be determined, by mean of the Witness experimenter tool.

As stated previously, the goal of the present problem is to reduce the

number of operators. After simulating the impact of the proposed so

lution on the studied picking process, the results indicate that there is a

very interesting improvement in the evaluation parameters (see

Table 4). The travel time decreases from 12.37 h to 1 h, reflecting a

91.6% improvement in the time lost in round trips between the "trolley

zone" and the storage racks. Furthermore, the delay to fill the 19

H

Trolleys zone

1 2 3 4 5 6 7 8 9 10

1

1

1

2

1

3

1

4

1

5

1

6

1

7

1

8

1

9

Rack

Trolley

Fig. 10. Witness model of the warehouse problem.

Table 4

Values of the evaluation parameters for the different situations.

Number of

operators

Delay to fill the 19

trolleys (h)

Total load for

operators (h)

Medium capacity

reserve (h)

Total waiting

time (h)

Total travel time of

operators (h)

Initial situation

7

4.44

30.25

3.68

0

12.37

Improved situation

7

3.22

20.73

5.04

1.95

1.00

Optimal solution obtained with optimization 4

4.98

19.42

3.14

0.64

1.00

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

134trolleys decreases from 4.44 h to 3.22 h. Thus, the number of operators

can be decreased since there is a margin of 1.78 h to reach the limit of

"the delay to fill the 19 trolleys" (which must not exceed 5 h). There

fore, by using the Witness 14 "experimenter" tool, the new system

configuration reduces the number of operators to 4 operators per shift.

Thus, the system gains 3 operators per shift, which equates to 9 op

erators for the 3 shifts.

4.3. Step 3: Use the design of experiments (DoE) method

In this step, two designs of experiments are realized in the Witness

experimenter tool, the design of experiments of the initial situation

(DoE1) (see Table 5) and design of experiments of the improved si

tuation (DoE2) (see Table 6), to determine the relation between the

input and output parameters of each system and to visualize the dif

ference between the two situations.

DoE1 and DoE2 are characterized by:

(1) The set of action parameters X = (X1 = sequence of picking,

X2 = number of available operators, X3 = type of trolleys).

(2) The set of evaluation parameters Y= (Y1 = delay to fill the 19

trolleys, Y2 = total load for operators, Y3 = medium capacity re

serve, Y4 = total waiting time, Y5 = total travel time of operators).

(3) The set of experiments E = (E1, ..., E19) for (DoE1) and

E = (E20, ..., E38) for (DoE2).

4.4. Step 4: Extract and formulate the generalized technical contradictions

of the problem

In this step, a step-by-step approach is followed to extract the best

system of contradictions corresponding to the studied problem. The

system of contradictions represents the entry point for the next step.

The evaluation parameters of the two designs of experiments should

be transformed into binary values, in order to clearly identify the ob

jectives achieved in each experiment. Table 7 combines the results of

DoE1 and DoE2, where an evaluation parameter takes the value "1" if it

is satisfied by the experiment Ei, and takes the value "0" otherwise.

From Table 7, it is seen that no solution can be found, because no

Table 5

Design of experiments of the initial situation (DoE 1).

Experiments X1: sequence of

picking

X2: number of

available operators

X3: type of

trolleys

Y1: delay to fill the

19 trolleys (h)

Y2: total load for

operators (h)

Y3: medium

capacity reserve

(h)

Y4: total

waiting time (h)

Y5: total travel time

of operators (h)

E1

random

1

fixed 30.15

30.15

0.00

0.00

12.37

E2

random

2

fixed 15.13

30.17

0.00

0.00

12.37

E3

random

3

fixed 10.15

30.19

0.00

0.00

12.37

E4

random

4

fixed 7.65

30.20

0.45

0.00

12.37

E5

random

5

fixed 6.15

30.21

1.96

0.00

12.37

E6

random

6

fixed 5.14

30.23

2.96

0.00

12.37

E7

random

7

fixed 4.44

30.25

3.68

0.00

12.37

E8

random

8

fixed 3.97

30.27

4.22

0.00

12.37

E9

random

9

fixed 3.61

30.26

4.64

0.00

12.36

E10 random

10

fixed 3.30

30.31

4.90

0.00

12.39

E11 random

11

fixed 3.09

30.33

5.24

0.00

12.37

E12 random

12

fixed 2.91

30.32

5.47

0.00

12.37

E13 random

13

fixed 2.76

30.34

5.67

0.00

12.37

E14 random

14

fixed 2.60

30.37

5.83

0.00

12.37

E15 random

15

fixed 2.51

30.38

5.97

0.00

12.37

E16 random

16

fixed 2.52

30.40

6.10

0.00

12.37

E17 random

17

fixed 2.50

30.41

6.21

0.00

12.37

E18 random

18

fixed 2.50

30.42

6.31

0.00

12.37

E19 random

19

fixed 2.51

30.41

6.40

0.00

12.37

Table 6

Design of experiments of the improved system (DoE 2).

Experiments X1: sequence of

picking

X2: number of

available operators

X3: type of

trolleys

Y1: delay to fill the

19 trolleys (h)

Y2: total load for

operators (h)

Y3: medium

capacity reserve

(h)

Y4: total

waiting time

(h)

Y5: total travel time

of operators (h)

E20 predefined 1

mobile 18.78

18.78

0.00

0.00

1.00

E21 predefined 2

mobile 9.53

18.94

0.00

0.16

1.00

E22 predefined 3

mobile 6.56

19.20

1.60

0.42

1.00

E23 predefined 4

mobile 4.98

19.42

3.14

0.64

1.00

E24 predefined 5

mobile 4.23

20.04

3.98

1.27

1.00

E25 predefined 6

mobile 3.61

20.48

4.57

1.70

1.00

E26 predefined 7

mobile 3.22

20.73

5.04

1.95

1.00

E27 predefined 8

mobile 2.88

21.15

5.36

2.37

1.00

E28 predefined 9

mobile 2.66

21.72

5.58

2.94

1.00

E29 predefined 10

mobile 2.48

21.85

5.82

3.07

1.00

E30 predefined 11

mobile 2.35

21.89

6.01

3.12

1.00

E31 predefined 12

mobile 2.20

22.09

6.15

3.32

1.00

E32 predefined 13

mobile 2.04

22.41

6.26

3.63

1.00

E33 predefined 14

mobile 1.91

23.04

6.35

4.27

1.00

E34 predefined 15

mobile 1.83

24.02

6.39

5.24

1.00

E35 predefined 16

mobile 1.82

24.92

6.44

6.14

1.00

E36 predefined 17

mobile 1.81

25.40

6.51

6.63

1.00

E37 predefined 18

mobile 1.81

25.96

6.56

7.18

1.00

E38 predefined 19

mobile 1.81

26.38

6.61

7.61

1.00

F.Z. Ben Moussa, et al.

Computers & Industrial Engineering 132 (2019) 124-140

135experiment Ei enables the satisfaction of all the evaluation parameters

Yj. To formulate a generalized system of contradictions, two sets of

evaluation parameters that are in contradiction should be determined.

Finding the best system of contradictions involves finding the existing

conflicts between non-dominated solutions. Thus, a dominance analysis

between the different experiments is performed to find the best non

dominated solutions. Indeed, based on the principle of Pareto dom

inance, one solution is better than another if it has at least as good

performance on all objectives and is more successful on at least one

objective. The calculation result of the dominance between the different

experiments is presented in the following table (Table 8). The binary

rank of a solution (or a set of solutions) is equal to the number