Technical Sessions A8 - G8

SESSION A8: Applied Operations Research

Malihe Aliasgari, Mohammad Kazemi, Reza Kia and Ahmad Ebrahimi

In this paper, a mathematical model with a weighted multi-objective function is formulated to find the machine cells and part families in a cell formation problem. Three considered objectives are minimizing the total number of exceptional elements and voids and total forming costs of cells. These objectives are combined with giving weights to them. Simultaneously minimizing both of exceptional elements and voids within the objective function makes a trade-off between inter-cell and intra-cell movements to decrease the dependence of cells and to increase the utilization of the machines. The proposed model has the advantage of finding the optimal number of cells which has not yet considered in the most studies implemented in the literature of cell formation problem. This feature provides the flexibility for the model to from cells in different numbers to reach the minimum sum of exceptional elements and voids. Several numerical test problems taken from the literature are performed to verify the performance of the proposed model in comparison to the previous models. The numerical examples show that obtained results are better than those in the literature in terms of the sum of exceptional elements and voids or grouping efficacy.

[160] Complexity and Heuristics for Multi Bipartite Complete Matching Vertex Interdiction Problem: Application to Robust Nurse Assignment
Pierre Laroche, Franc Marchetti, Sébastien Martin, Zsuzsanna Roka and Mengli Zheng

In this paper, we consider the robust nurse assignment problem on several time slots. For one time slot, this consists in finding the maximum number of absences still permitting an optimal treatment of patients. This problem can be solved in polynomial time and several results are given in [7]. We focus specifically on the several time slots version, the goal is to find a distribution of nurses on all time slots, ensuring that the minimum number of possible absences on each time slot is maximum. We show the NP-completeness of the problem on several time slots, we give a greedy algorithm with some improvement and a genetic algorithm to solve this problem.

[164] New model for the robust vehicle routing problem with uncertain travel time
Lei Wu, Mhand Hifi and Hiba Bederina

In this article, we implement a new robust criterion for the robust vehicle routing problem with uncertainty on the travel cost. In order to simulate the uncertain events, we propose an integer linear program which is based on the discrete scenario optimization approach. The objective is to find a solution which performs a better behavior on a majority of scenarios. For this purpose, we propose a new robust criterion which is referenced by Maximizing the Number of Scenarios Qualified by the Worst (shortly MNSQW). In order to highlight the performance of the proposed approach, MNSQW is evaluated on a set of 480 instances generated from Solomon's benchmark instances. The solutions provided by MNSQW are compared with the other classical robust criteria. The numerical results show that MNSQW is able to produce the robust solutions in a majority of cases.

[28] Competitive Facility Location Problem with foresight Considering Service Distance Limitation
Xia Mingfei, Qi Mingyao, Zhang Ying and Miao Lixin

This paper presents a practical bilevel nonlinear integer programming model for the competitive facility location problem with foresight. The objective of the developed model is to maximize the leader’s market share while predicting the follower’s response. In the classical competitive facility location model, it is assumed that the facility competes for all customers no matter how far they are. Although widely applied the model is, the flaws are obvious. This paper considers the service distance constraint and propose a new customer behavior, which is more accurate than the existing models. For solving the model, a two-stage hybrid tabu search algorithm is proposed. To illustrate the effectiveness of the proposed algorithm, a set of randomly generated instances are presented and analyzed statistically. The results indicate that the proposed algorithm provides an effective means to solve the problems and the service distance is proved to be a significant factor in the model.

SESSION B8: Cloud Computing, Future Generation Computer Systems for High Performance Parallel, Distributed, and Network-Based Processing

[287] WSN localization scheme based on Received Signal Strength Indicator (RSSI) for ZigBee Networks
Monaem Idoudi, Jorge Cabral, El-Bay Bourennane and Khaled Grayaa

Wireless Sensor Networks (WSNs) have diverse application domains such as smart home, smart care, industrial, etc. We present a WSN system based on the ZigBee technology (IEEE 802.15.4) in Smart Home. In our paper we interest to ZigBee protocol is often used in medical Rehabilitation, which is a relatively new concept involving wireless transmission of data from the sensors attached to a patient to a distant monitoring station. There is no standardized topology managing the current networks, therefore, we will compare and evaluate the performance the mobility of nodes for star topologies in different scenarios to determine which is the most suitable in a typical hospital environment. The study of performance is based on measurements of the Received Signal Strength Indicator (RSSI) in different position in the Room of test using STM32F04 board. The analysis will be performed in OPNET comparing characteristics such as transmission efficiency and network delay.

Lijun Sheng and Xiaochuan Shi

In Intelligent Transportation Systems, Intelligent Traffic Management Cloud Platform can be used to improve system integration and compatibility. The service model of Intelligent Traffic Management Cloud Platform includes client, cloud and pipe, which is a core part and divided into Core Backbone Network and Road Access Network.The main work to construct the cloud platform is to construct Road Access Network, whose computational model is a weighted graph in which weight of edge is to be determined. After revising ETX,ETT,ETTI and other parameters, RETTI, which takes into account not only traditional factors such as delay but also channel interference and node load, is considered as the optimal parameter as the weight of edge.

[274] Aid to Design and Reconfiguration of the MPSOC Architectures
Naija Mohamed and Ben Ahmed Samir

Reduce the development complexity of MPSoC embedded architectures has become a major concern for designers of these systems. This current research is interested in the design of these latters by raising the abstraction level of the specification through a model-driven approach. It takes advantage of the different concepts provided by MARTE profile to model separately the functional part of the application, the available hardware architecture and the association between them. Our study focuses specifically on the temporal analysis of the system by using the theory of scheduling to test the feasibility of the studied system. The result of the analysis can then be used for the reconfiguration of the high level modelled system.

SESSION C8: Complex Systems & Applications

Nader Nematpour and Kaveh M. Cyrus

The Many Iranian Corporations spite of having sufficient available resources and written challenging future events; don’t use their resources in proper and efficient ways. Therefore, are not able to achieve their missions and goals. To face this major problem, it is required to have a productive management system interacting with innovation promotion (in product/service/working process) and stakeholder benefits. Such Methodology will result on sustainable productivity and survival situation for organization even in turbulent environment.

Shanshan Xu, Dongqin Feng and Jianfei Chen

For lack of identity authentication mechanism in industrial control field, this paper established an authentication system for industrial control network, and designed an authentication scheme based on dynamic password as well as random key agreement to achieve critical components identification in control network. The scheme contains functions of authentication, key update and heartbeat certification to ensure the security, reliability and availability of communication in control system. Finally, a strand space model of the scheme was built to perform security analysis and practical experiments were taken to demonstrate the execution efficiency.

Fatma Lajmi, Achraf Jabeur Telmoudi and Lotfi Nabli

Monitoring plays a major role in safely operating production systems. This paper proposes an approach to help decision making in maintenance strategy based on the FMEA tools and Bayesian networks. In this paper, we present our approach to exploit these essential characteristics, focusing on monitoring the capabilities of a complex system using Bayesian networks.The proposed approach is rather important in maintenance management. We use the FMEA (Failure Mode Analysis and Effects) to identify critical components of a production line. Based on result of FMEA the Bayesian network model is developed.

SESSION D8: Applied Operations Research

Fadel Abdallah, Camel Tanougast, Imed Kacem, Camille Diou and Daniel Singer

The problem of optimizing the computational resources is still open. This article studies the scheduling problem of heterogeneous architecture CPU/FPGA with heterogeneous communication delays. In this context, two approaches based on genetic algorithms are proposed and implemented. Their main objective is to run in the MPSoC an application described in the data flow graph form in order to minimize the scheduling length Cmax (Makespan). Results have demonstrated that both proposed approaches offer an approximate or optimal solution for the scheduling task problem as well as to obtain better solving time compared to similar works.

[302] On the Knapsack Problem with Overlopping Properties
Aristide Grange, Imed Kacem, Karine Laurent, Sébastien Martin

In this paper, we consider a knapsack problem under merging objects'. This problem is a variant of the well-known Knapsack Problem where the size of the items in a knapsack is not the sum of all items taken. We show the NP-completeness in strong sense of this problem. We propose some methods to solve this problem with heuristic and exact methods. We propose several bounds to improve all exact methods.

[298] On the Visualization of the Information in Social Numerical Networks
Youcef Abdelsadek, Kamel Chelghoum, Francine Herrmann, Imed Kacem, Benoît Otjacques

The interest given to the community detection algorithms in graphs is continuously increasing. The objective of such algorithms is to reveal the semantic of communities underlying to the graph structure. In this paper we propose a community detection algorithm in weighted graphs. The proposed algorithm is based on the weighted maximum triangle packing to build the skeleton of communities at its first step. Thereafter, the algorithm compares the intra-community weight and inter-community weight between communities allowing dominant communities to gain in size. Furthermore, we introduce our application to investigate the detected communities. NLCOMS is a visual interactive application for community detection based on node-link representation of graphs and circle packing for the detected communities. Finally, the algorithm is assessed on real-world data of the info-RSN ANR project.

[295] Genetic Algorithm for Scheduling Surgical Units
Rania Ayachi, Imed Kacem, Myriam Sahnoune, Achraf Telmoudi

This paper considers an important scheduling problem in surgical units with the aim of minimizing the impact of simultaneous arrivals of emergent patients. We propose a genetic algorithm to solve the considered problem. Numerical experiments have been conducted and the obtained results show encouraging performances.

[133] Two Developed Models for a Green Vehicle Routing Problem with Safety and Social Concerns
Mahdi Bashiri and Arghavan Sharafti

Green vehicle routing problem (GVRP) is one of recent variants of vehicle routing problem (VRP), dealing with environmental aspects of distribution systems. In the literature, economic aspect has been often used, while safety and social concerns have less proportion of studies. In this paper, we present two developed mixed integer-programming models for GVRP with social and safety concerns. Different numerical analyses have performed to evaluate suggested models and investigate the influence of several key factors in them. The results confirm that the proposed model has more social benefits such as balanced tours and less customer waiting time than the classic GVRP.

SESSION E8: Hub-and-Spoke operations in land and maritime transport

Khadidja Yachba and Karim Bouamrane

This article addresses the problem of containers storage in a container terminal. Container terminals constitute essential intermodal interfaces for the global transport network. The optimum location for a container in terminal is very important for port operators as it directly impact the overall performance of container terminal and play an important role in maximizing throughput as a results of higher efficiency. In our work, we propose an approach of container placement problem through the description of a model to simulate, solve and optimize available storage space to handle departures and arrivals of full containers in a seaport. In other words, a model that minimizes the total number of unnecessary movements while respecting dynamics of environment with respect to the space and time.

Ines Rekik, Sabeur Elkosantini and Habib Chabchoub

Container Stacking Problem (CSP) is of great importance in a seaport terminal with regards to the increasing number of containers to be handled. Numerous research studies related to CSP have been published. This paper presents a comprehensive review of the literature on this problem. Firstly, we present existing studies related to CSP with regards used control strategies and also used Key Performance Indicators developed to maintain the competitiveness of a storage yard. Secondly, we analyze the variants of the CSP and that are related to either static or dynamic considerations. The used approaches to solve CSP including optimization methods and artificial intelligence approaches are then distinguished. Several research trends are finally identified in the conclusion.

Naima Belayachi, Shahin Gelareh and Karim Bouamrane

Maritime transport represents as the preferred mode for the global trade. Indeed, containerization is regarded as one of the most remarkable improvements in maritime transport. The containers are rented (or owned) by shipping companies. However, these companies encounter excess of empty container at some ports and lack others in different ports while meeting customer demands. This lack of empty containers at certain ports of Maritime Transport Network (MTN) is simply a consequence of the imbalance in the distribution of their containers through the network because the containers are sent and did not return thereafter. The objective of this work is to solve the problem of imbalance in the distribution container level to look for empty containers and meet customer demands. To this end, we propose a model for representing the MTN, and to ensure a balanced distribution of the containers. This paper presents a model of an MTN to simulate customer demand management and the transfer of full containers and the search for empty containers. Our approach is based on a heuristic method by neighborhood. The principle of this method is to conduct a heuristic local search starting from an initial solution and moving to other neighbors, in order to allow the search for empty containers if a port do not dispose in its stock to satisfy customer demand from origin to destination.

SESSION F8: IE Research, Education and Practice in China and East Asia

[137] New demands for Lean Production in the context of customized production and smart manufacturing
Ting Qu

The core concept of lean production is to pursue waste reduction and cost saving. Mass production establishes working standards for the front-line workers to realize leanness in the behavioral level. Small-batch and high-variety production triggered the emergence of Just-in-Time (JIT) methods to pursue waste elimination through process synchronization, i.e. Toyota synchronize the material demand and supply between adjacent stages to eliminate in-process inventory. Nowadays, on the one hand, the rapid involvement of automatic machines and robots are largely reducing the human’s behavioral influence to the production efficiency. On the other hand, customized or personalized product, which takes the form of much smaller-batch and higher-variety product, have put higher requirements to the flexible organization of production system and the smart management/control of production process. Therefore, the following three considerations have become new demands of lean production in the new era. First, how to appropriately organize internal and external resources toward specific production requirement to realize structural leanness. Second, how to make smart and adaptive process plan based on the dynamically involved resources to realized process leanness. Third, how to accurately sense the real-time statues of both resource and progress to realize management leanness. According to the over ten-year practical collaboration with the manufacturers in China’s PRD region, which is referred to as World’s Factory, we found that although few of them have truly realized or even begun the Industrial 4.0 implementation, but some leading manufacturers have already launched product customization strategy and are actively exploring smart and lean solutions. A smart production-logistics (PL) synchronization system which is recently developed by the author’s project team for a large-size paint making company to deal with both the high-variety and small-batch product demand and the dynamic production environment will be chosen for examination. Specifically, the cloud-based group-level PL resource management concept, the Internet-of-Things based real-time PL process management, as well as the real-time data-driven smart PL synchronization mechanism will be introduced to demonstrate its lean production concept.

[300] Lean Six Sigma and Management Innovation in China
Zhen He

Juhong Gao, Rui Wang and Haiyan Wang

[237] Lean Production Practice in China——a Case in Coal Industry
Feng Wu

Lean production (LP) has been introduced to China over 15 years, except manufacturing enterprises, LP has been practiced in various industries by State-Owned Assets Supervision and Administration in China during recently years. A case on LP in the largest enterprise in coal industry of China is presented which includes question investigation, OEE measurement, system analysis, alternatives and outcome. The case focuses on the productive promotion. A management implicit is proposed finally.

SESSION G8: Mining and classification models for Biomedical data or image analysis

[90] Image Data Monitoring Using Nonparametric Profile Monitoring Techniques
Mehdi Koosha and Rassoul Noorossana

Image data plays an important role in new emerging real- life problems because of their tremendous superiorities in comparison with classic data types. Each image can provide millions of data points from product in a short time. Handling this data volume requires new methods. This paper deals with proposing a new statistical method based on nonparametric profile monitoring techniques to monitor image data for statistical process control purposes. Performance of the proposed method is evaluated using simulation study.

Ahmad Chaddad, Ahmed Bouridane and Camel Tanougast

Cancer causes heterogeneity in the biological structures of cells. Texture analysis of continuum cancer can characterize these structures and provide an accurate cancer diagnosis and grading. This paper introduces an effective texture analysis of colorectal cancer to characterize the biological structures in three abnormal cell (ABC) types (benign hyperplasia (BH), intraepithelial neoplasia (IN) and carcinoma (Ca)) for tissue quantification. We segmented the abnormal cells for texturizing by a Laplacian of Gaussian filter which generates three texture types namely fine, medium and coarse. Texture was then quantified for each group of the three abnormal cells. Working with multispectral of tissue images, our preliminary experiments demonstrate that the proposed texture model leads to over 81.07 % of classification accuracy using a linear discriminant analysis. This demonstrates the feasibility of continuum cancer assessment by using non-invasive technique by the texture feature extraction.

Hawraa Haj-Hassan, Ahmad Chaddad, Camel Tanougast and Youssef Harkouss

Upcoming the past decades, image segmentation became an urgent need due to its significance in the biomedical field. This paper organized segmentation techniques into three, namely: level set, edge-based and a hybrid between these methods. These techniques have been compared, and hybrid method was demonstrated with a high performance capability to detect irregular shapes. This study showed the advantage of deformable technique to segment abnormal cells with Dice similarity value over 80%.