Technical Sessions A5 - E5

SESSION A5: Healthcare engineering and management


[85] A HEURISTIC FOR DEFINITION OF SHIFTS IN AN EMERGENCY DEPARTMENT
Karim Ghanes, Agapitos Diakogiannis, Oualid Jouini, Zied Jemai and Mathias Wargon

We propose a model that combines simulation optimization and linear programming in order to define the shift pattern that best match the arrival pattern of patients in an Emergency Department (ED). The simulation model supplies the linear programming with the staffing levels that secure the performance of the ED, expressed in terms of the average Length of Stay (LOS) of patients. The linear model determines the shift scheduling of all employees with the use of the minimum cost, including several practical constraints as experienced in practice. The model includes also a heuristic that helps ED managers to find a solution that satisfies budget restrictions. This heuristic proposes an efficient manner of reducing the staffing cost, by decreasing the staffing levels in a way that affects the performance as less as possible. The application of the developed method leads to a reduction of 8.9% in the ED average LOS with the use of the same staffing budget.

[150] EVALUATING THE ADAPTATION OF A DWELLING TO A DISABLED PERSON: A GRAPH-BASED APPROACH
Yahya Bouzoubaa, Christian Minich, Anass Nagih and Pruski Alain

This paper presents a methodology and a computer tool to check if a dwelling meets both the needs of a disabled future occupant and the wishes he expressed about the layout of his dwelling. The major steps of the methodology are a translation of the needs and wishes into predefined criteria that can be handled by the software and the modelling of the dwelling by a graph that captures both the layout of the dwelling and the capabilities of its user. Together, these two features allow to take various types of physical disabilities into account and offer the occupant the possibility of formulating a wide range of demands. This evaluation module is one of the two main components of a decision support tool aiming at helping human experts to adapt the layout of a dwelling to its future occupant or the layout of a building to its future users.

[174] A PROBABILISTIC CHOICE MODEL FOR LOCATION OF PREVENTIVE HEALTH CARE FACILITIES WITH CONSTRAINED QUEUING SYSTEM
Mehdi Seifbarghy and Leila Rostamian

The importance of preventive health care programs is very high for the organizations, which are responsible for. However, the problem with that is the low participation rate of people. This paper presents a model for location of preventive healthcare facilities and uses the spatial interaction models to describe probabilistic behaviour of clients in patronizing established facilities and increasing level of participation. The proposed model is a nonlinear integer programming (NLIP) for which we have utilized genetic algorithm and simulated annealing in order to solve the problem. A number of numerical examples are designed and solved in order to assess the performance of the solution algorithms.

[203] HEALTHCARE LOGISTIC NETWORK MODELS FOR SHARED RESOURCE MANAGEMENT
Imene Elhachfi Essoussi and Pierre Ladet

In this work, we propose a simulation–based optimization approach that addresses the problem of the design of collaborative healthcare logistic networks through the re-thinking of the inventory management process and the resources sharing rules. We especially analyze the gains from medical supplies pooling within a group of nearby hospitals through the instrumentation of the inventory centralization scheme and the rules for lateral inventory sharing such as the emergency lateral transshipment. Denis Cornaz, Youcef Magnouche, Ali Ridha Mahjoub and Sébastien Martin. The multi-terminal vertex separator problem: polyhedral analysis and branch-and-cut Abstract: Let G = (V _ T, E) be an undirected graph such that V is a set of vertices, E a set of edges and T a set of terminal vertices. The Multi-terminal vertex separator problem consists in partitioning V into k + 1 subsets {S, V_1 , ... , V_k } minimizing the size of S and such that there is no edge between two subsets V_i and V_j and each subset V_i contains exactly one terminal. Set S is called a separator. In this paper, we show that this problem is NP-complete. We discuss the problem from a polyhedral point of view. We describe some valid inequalities and characterize when they define facets. Using this we develop a Branch-and-Cut algorithm.

SESSION B5: Manufacturing


[66] INVESTIGATION OF VISUAL MANAGEMENT CASES IN CONSTRUCTION BY AN ANALYTICAL FRAMEWORK FROM MANUFACTURING
Koichi Murata, Kazuki Nakajima, Koichi Kinoshita, Algan Tezel, Lauri Koskela, Patricia Tzortzopoulos and Hiroshi Katayama

Along with the progress of globalization, speed and efficiency have become more critical for any industry than ever before. In this sense, the concept and methods of lean management, promoting these performances, have been deployed from manufacturing, its origin industry, to other industries. This paper deals with this management style in the construction industry, called lean construction. In particular, visual management (VM) as one effective tool in this scheme is focused on. A number of VM cases, 306 in total, was collected from both construction and manufacturing sites and investigated by the so-called 5W1H analytical framework developed in the manufacturing industry. Obtained results suggest that the VM cases in construction have common attributes such as purpose and location, target to attain, users’ attributes, timing to use and elemental technologies for case development. A comparison analysis of the VM cases from construction and those from manufacturing was also carried out, for a mutual transfer of this technology between these industries.

[149] A STUDY ON DESIGN OF VISUAL MANAGEMENT SYSTEM FOR IMPROVING ITS PERFORMANCE
Koichi Murata

This paper discusses the design of a visual management (VM) system. It is recognized as one of the recent trend management systems and there are found many relevant cases in factories around the world. However the study of how to design a VM system is not enough in the academic field. If non-systematic development of a VM system continues, low performance cases will increase. Consequently, not only VM but also the capability of the problem-solving in manufacturing industry begins to decline. Therefore in this paper, the basic mechanism of a VM system is proposed. And then, the improvement of its capability is performed through the analysis of three cases by the proposed mechanism for its systematic design. The obtained result of this paper is the procedure for designing a VM system and five points utilized in the procedure; 1) comprehensibility, 2) non-ordinarily, 3) continuity, 4) inducibility and 5) immediacy.

[193] REAL-TIME PRODUCTION PERFORMANCE ANALYSIS AND EXCEPTION DIAGNOSIS SYSTEM FOR THE INTERNET OF MANUFACTURING THINGS
Yingfeng Zhang and Wenbo Wang

In order to improve the capability for online supervision and prompt exception management, this paper proposes a framework for the real-time production performance analysis and exception diagnosis method by extending IoT (Internet of Things) technology to the field of manufacturing system. Under the Framework, the real-time primitive manufacturing events which occurred at different manufacturing things such as operators, machines, pallets, WIPs and so forth can be easily sensed. Then, the critical production performance can be obtained based on the Timed Coloured Petri Net model, where the token can change with the real-time status of smart objects. At last, the exception event and its cause are diagnosed based on Decision Tree technology.

[196] METRICS AND METHODS IN LEANNESS ASSESSMENT: A LITERATURE REVIEW
Pinar Yurdaer and Huseyin Selcuk Kilic

Lean production which mainly targets the elimination of waste and increase in customer satisfaction requires steps to be followed for implementation. As the literature suggests, the first step of lean implementation is leanness assessment. The implementation process starts after the existing situation is evaluated with respect to the lean indicators via a proper method. The two important factors in leanness assessment are the lean metrics and the methods that are used. There are various studies using quantitative and qualitative lean metrics and different methods. In this study, numerous papers are reviewed and the lean metrics and methods are identified and analysed with respect to their frequencies. Moreover, the fields that need further research areas are stated.

SESSION C5: Heuristics and Approximation Algorithms for Scheduling Problems


[86] EXTENDED MODEL FORMULATIONS FOR CELL LOADING AND FAMILY SCHEDULING PROBLEM WITH INDIVIDUAL DUE DATES TO MINIMISE MAXIMUM TARDINESS
Cemalettin Ozturk and Gursel Suer

In this paper, we present two mathematical model formulations for cell loading and family scheduling problem with individual due dates to minimize maximum tardiness. The problem and an initial model formulation along with a genetic algorithm were presented in a recent paper where sequence independent setup times between jobs of different families were assumed. In this study, first, we develop a tighter model formulation for the same problem. Then, we extend the tighter model formulation to manufacturing cells in which sequence dependent setup times exist. We also provide numerical examples to demonstrate application of the model formulations.

[104] JOINT RESOLUTION OF A JOB-SHOP SCHEDULING PROBLEM WITH A ROUTING PROBLEM
Murat Afsar, Philippe Lacomme, Libo Ren, Caroline Prodhon and Daniele Vigo

This paper focuses on the integration of distribution routing issues into the production scheduling in order to minimize the makespan. The overall problem consists in simultaneously scheduling a set of jobs in a Job-Shop framework, and routing the completed jobs between the different phases of the production environment with a homogenous fleet of vehicles. The routing must take into account both the loading capacity of the vehicles, and for each operation the ready-time at a given machine and the due date at the following one. Job-dependant processing time, delivery time windows, and destination are addressed into a GRASP_ELS approach solving the whole problem into an integrated framework. The framework has been tested on Job-Shop instances and obtains results very close to solutions provided by dedicated methods. A new set of instances with up to 100 operations to be scheduled is introduced and new solutions are provided. The result demonstrated the interest to consider the routing issue during the computation of the makespan.

[111] GENETIC ALGORITHMS FOR SCHEDULING A NO-WAIT TWO-MACHINE FLOW-SHOP TO MINIMIZE THE MAKESPAN UNDER NON-AVAKLABIMI\Y CONSTRAINTS
Faten Ben Chihaoui, Atidel B.Hadj Alouane and Imed Kacem

A two-machine no-wait flow-shop scheduling problem to minimize the makespan is considered in this paper. Every machine is subject to one non-availability constraint. The non-availability intervals of the machines overlap and they are known in advance. The jobs have different release dates. Genetic algorithms are developed to solve the problem. Computational experiments are carried out on a large set of instances. The obtained results show the good performance of the proposed method.

[135] Stability measures for a stochastic multi-level job scheduling problem with daily working hour
Byung Jun Joo, Sang-Oh Shim, Tay Jin Chua and Tian Xiang Cai

This paper focuses on a single machine multi-level job scheduling problem under stochastic processing times and daily working hour. For a bi-objective of minimizing the total weighted tardiness of customer orders and the schedule deviation of the realized schedule from the baseline schedule, three stability surrogate measures for the schedule deviation and a metaheuristic scheduling algorithm are developed to cope with the processing time uncertainty as well as the daily working hour. Result of the computational experiments with randomly generated problem instances is provided to evaluate the performance of three surrogate measures.

SESSION D5: Applied Operations Research


[277] A MATHEURISTIC FOR SOLVING AN INDUSTRIAL PROBLEM OF CUTTING MOUSSE BLOCKS: A PRACTICAL TUNISIAN CASE
Mariem Baazaoui, Hichem Kamoun and Saïd Hanafi

The aim of this paper is to present a method for optimizing the cutting of mousse blocks in Tunisian industrial company. This industrial problem is closely related to the three-dimensional multiple-bin-size bin packing problem. This problem consists in identifying an optimal cutting of the material during the manufacturing process. This cutting must satisfy certain geometric constraints and aims to reduce losses and costs. The current work proposes a matheuristic method which combines mathematical programming techniques and heuristics for this cutting mousse blocks problem by considering a daily demand and a variety type of items. The preliminary results show its strength in reduced trim loss through its results.

[246] IMPROVING THE COMPLEXITIES OF FPTAS FOR SCHEDULING AND ROUTING PROBLEMS
Eugene Levner, Amir Elalouf and Imed Kacem

We suggest a general technique for improving the complexity of fully polynomial time approximation schemes (FPTAS) for several combinatorial problems. Using the dynamic approximate binary search, we first obtain an improved FPTAS for a problem of scheduling temperature-dependent operations on a single machine. Then, we show that the suggested technique can be used for improving the worst-case complexities of earlier FPTASs for several other routing and scheduling problems.

[252] Production order quantity under uncertainties and forecasts update: optimal control quantity using stochastic dynamic programming
Maxime Claisse, Zied Jemai and Chengbin Chu

In this paper, we consider a single level single item production control quantity problem where the demand is represented by a stochastic and non-stationary variable. The update of the information is taken into account through a rolling horizon framework: the demand of a specific period $t$ depends on all past information on hand at this time, and is updated through the dynamics of the forecast method used. We formulate the problem as a stochastic dynamic programming problem, and give an explicit optimal solution for a case where the production lead-time is reduced to zero. We finally compare the optimal results with classical production planning methods through a simulation.

SESSION E5: Manufacturing


[15] CREATING INDEX FOR SUSTAINABILITY/SUSTAINABLE DEVELOPMENT IN AN INDUSTRIAL ESTATE
Ibrahim Garbie

Creating an index for sustainability/sustainable development (S/SD) is important and urgent to be formed although there are a lot of researchers from academicians and industrialists talking about the S/SD. The main challenge is estimating the S/SD index among different industries in one industrial region. The main objective of this paper is to present a framework to model and assess an index of S/SD within an industrial estate with a group of different industries. Sustainability indexes for each assessed value (indicators, aspects, dimensions, industry type and whole region) will be normalized, evaluated and weighted. An illustrative example is used to illustrate the proposed framework.

[47] DEADLOCK CONTROL DESIGN AND PLC IMPLEMENTATION OF MANUFACTURING SYSTEMS
Husam Kaid, Abdulaziz El-Tamimi, Emad Abouel Nasr, Abdulrahman Al-Ahmari and Zhiwu Li

Petri nets are an effective way to model, analyze, and control deadlocks in automated manufacturing systems (AMSs). The need for an effective design tool to design a control system and convert Petri net to ladder diagram for PLC implementation becomes increasingly more important. Therefore, the objective of this paper is to design behaviorally optimal liveness-enforcing supervisor for automated manufacturing system and their ladder diagram for programmable logic controller (PLC) implementation. To do this, first, deadlock prevention method (elementary siphons control method) has been used to design a supervisor to prevent the deadlock in AMSs. Next, a token passing logic (TPL) technique has used to convert the resulted controlled model into ladder diagrams (LDs) for implementation. Finally, the presented methodology can be applied for multiproduct and sizes of AMSs and provided a more effective for PLC implementation.

[51] ROBUST OPTIMIZATION TO DESIGN A DYNAMIC CELLULAR MANUFACTURING SYSTEM INTEGRATING GROUP LAYOUT AND WORKER ASSIGNMENT
Reza Kia, Mohsen Lashgari and Fariborz Jolai

In this paper, robust optimization of a mathematical model integrating cell formation, intracellular/intercellular layout and operator assignment is implemented in a dynamic environment. The important concern of the employed robust methodology is to obtain an optimal cellular manufacturing design that is robust with regard to data uncertainties in part demand and processing time. Since intracellular/intercellular layout and operator assignment are interrelated decisions, they should be determined simultaneously to reach an optimal solution. Hence, an integrated model is formulated to incorporate several design features including uncertain processing time, uncertain part demand, operation sequence, intracellular/intercellular layout, multi-rows layout of equal area facilities, flexible reconfiguration, hiring/firing and training of operators, operator capacity and cell size limits. The objective of the proposed model is to minimize total costs of intra/intercellular material handling, relocation and installation/uninstallation of machines, and operator related costs (i.e., hiring/firing, training and salary). An illustrative numerical example is solved to demonstrate the robust model validity.

SESSION F5: Polyhedra  and  Combinatorial  Optimization


[45] A pseudo-polynomial size formulation for 2-stage 2-dimensional knapsack problems
Fabio Furini and Enrico Malaguti

Two dimensional cutting problems are about obtaining a set of rectangular items from a set of rectangular stock pieces and are of great relevance in industry, whenever a sheet of wood, metal or other material has to be cut. In this paper, we consider the 2-stage two-dimensional knapsack (2TDK) problem which requires finding the maximum profit subset of rectangular items obtainable through 2-stage guillotine cuts in a rectangular panel. We propose a formulation having a pseudo-polynomial number of variables and constraints which can still be enumerated for the instances present in the literature. We compare the proposed formulation with the previous best known polynomial size one. Extensive computational experiments show that the new model is characterized by a stronger linear programming relaxation and can be effectively solved with a general-purpose MIP solver.

[93] THE K-NODE CONNECTED SUBGRAPH PROBLEM: FORMULATION, POLYHEDRA AND BRANCH-AND-CUT
Meriem Mahjoub, Ibrahima Diarrassouba and A. Ridha Mahjoub

In this paper we consider the k-node connected subgraph problem. We propose a new integer linear programming formulation for the problem and investigate the associated polytope. We introduce further classes of valid inequalities and discuss their facial aspect. We also devise separation routines. Using those results, we devise a Branch-and-Cut algorithm along with some preliminary computational results.

[98] ENERGY SAVING IN CONTENT-ORIENTED NETWORKS
Nicaise Choungmo Fofack, Ali Ridha Mahjoub, Mohamed Yassine Naghmouchi and Nancy Perrot

By allowing in-network caching, content-oriented networks may significantly decrease the network congestion, shorten the access delays, and reduce latency when delivering contents. On the other hand, a massive deployment of caches may subsequently increase the operational expenditures (OPEX), and particularly the energy bill of telecommunication operators. In this paper, we address the energy saving problem in content-oriented networks. This consists in determining which caches and which links could be switched off to minimize energy consumption in such a way that all demands are met while respecting capacity constraints. We propose a novel Mixed Integer Linear Programming (MILP) Formulation of the problem to solve the related object caching and traffic routing problem on arbitrary graph-based network topologies. We use CPLEX to solve our model to optimality. Then, we assess several network performance metrics. After all, we develop a routing on shortest path-based heuristic in order to compare our solutions with those given by the standard shortest path-based routing. Finally, we discuss the numerical results. We show that: 1) The metrics of interest provide additional insights on the impact and/or gain of introducing energy-aware caches in a real telecommunication network; 2) The benefits of our model compared to a routing on shortest path-based model: 38.72 % of energy saving is reached using our MILP model.

[205] The multi-terminal vertex separator problem: polyhedral analysis and branch-and-cut
Denis Cornaz, Youcef Magnouche, Ali Ridha Mahjoub and Sébastien Martin

Let G = (V _ T, E) be an undirected graph such that V is a set of vertices, E a set of edges and T a set of terminal vertices. The Multi-terminal vertex separator problem consists in partitioning V into k + 1 subsets {S, V_1 , ... , V_k } minimizing the size of S and such that there is no edge between two subsets V_i and V_j and each subset V_i contains exactly one terminal. Set S is called a separator. In this paper, we show that this problem is NP-complete. We discuss the problem from a polyhedral point of view. We describe some valid inequalities and characterize when they define facets. Using this we develop a Branch-and-Cut algorithm.

SESSION G5: Virtual product development


[138] Product variety and reconfigurable process plan modeling for new manufacturing paradigms
Qing Xia, Alain Etienne, Jean-Yves Dantan and Ali Siadat

Mass customization and co-evolution of product, process and production system are two prevalent manufacturing paradigms which can help manufacturers cope with the manufacturing complexities introduced by product variety and manufacturing dynamics. Product variety modeling and reconfigurable process planning are two interdependent key techniques for the implementation of these two paradigms. Focusing on the current two main research issues of reconfigurable process planning, a feature-based product variety model structured by applying three logical operators is proposed and a set-based method to represent the design specifications of a feature family is presented. In addition, three levels of reconfigurable process plans are discussed and a model for reconfigurable machining operation plan is defined. The proposed concepts and models are illustrated and justified with the example of an oil pump body family. This example shows that the proposed approaches can provide the flexibility to handle the changes on product design and have the effectiveness for the implementation of design for manufacturing.

[244] A FUNCTION ANALYSIS APPROACH TO REDUCE COMPLEXITY IN PRODUCT DESIGN
Roozbeh Babaeizadeh Malmiry, Jean-Yves Dantan, Jérôme Pailhès and Jean-François Antoine

During product design process, designers often deal with complexity which is caused by variety and number of elements in the process as well as the dependencies among them. This paper proposes the modelling language of Characteristics-Properties Modelling (CPM) as a framework to model the product and later to analyze it. CPM is a design approach based on the behavior of the system. It enables the designer to dimension the product in each step of design process. Moreover, by using CPM, one can analyze the output of the system according to these dimensions. In addition, the model of ‘CTOC’, which is a systematic modelling tool, is used to help the designer in the process of design. It decomposes the system based on the functioning of the product. The proposition of this paper is the interoperability of CTOC and CPM to reduce complexity in design process. The paper proposes an approach for the coherence of CTOC with CPM. The result is a quantitative multi-physical approach to model and analyze the product. By means of this interoperability, the designer will be able to improve the product performance by means of the modifications such as dimensions, tolerances and/or material. Thus, by decoupling the element of design using the proposed approach, the designer is able to reduce complexity in design process.

[262] THE IMPACT OF IMMERSIVE VIRTUAL REALITY ON COLLABORATION IN DESIGN MEETINGS: A COMPARATIVE STUDY
Ahmad Al Khatib, Damien Fleche, Morad Mahdjoub and Jean-Claude Sagot

Collaboration is needed between different actors into product design process. This could be a problematic especially when design teams are from various backgrounds and have different expertises. Nowadays virtual prototypes based on immersive Virtual Reality (VR) technology are used for several purposes by design teams throughout all product design process. For instance, they are used as intermediary objects that facilitate social interactions and collaboration in collaborative design meetings. In this paper, we investigate the benefits of VR technology to improve collaboration between design actors in collaborative design meetings. To this end, we conduct an experiment to compare VR condition with traditional CAD condition. Results show that VR condition improves collaboration for several aspects like acquisition of new knowledge and perceived quality of communications and group motivation. However, the results don’t show any effects on other aspects like the perceived effectiveness of design reviews and the effective participation in design review.