ICAOR’12 ABSTRACTS

4TH INTERNATIONAL CONFERENCE ON APPLIED OPERATIONAL RESEARCH

25-27 JULY 2012, BANGKOK, THAILAND


OPERATIONS RESEARCH AND DYNAMIC PROJECT SCHEDULING: WHEN RESEARCH MEETS PRACTICE

Mario Vanhoucke

Faculty of Economics and Business Administration, Ghent University, Gent, Belgium

Operations and Technology Management Centre, Vlerick Leuven Gent Management School, Gent, Belgium Department of Management Science and Innovation, University College London, London, UK

Abstract. In this study, an overview is given of recent developments in the dynamic project scheduling literature. Both resource-constrained project scheduling and project risk analysis have been widely investigated in the academic literature as useful tools to control projects in progress. Project control has recently received a renewed research attention since the revival of academic publications on earned value management. The integration of academic results in a novel software tool will be discussed from a dynamic scheduling point of view and some practical implications are illustrated. The software tool makes use of state-of-the-art algorithms discussed in the literature and can be used for both commercial and academic purposes. It will be shown that the algorithms implemented in this tool are based on state-of-the-art research results that will be continuously improved by new research results. Therefore, the tool will also be used as a research engine to stimulate future researchers to develop improved algorithms for project scheduling, risk analysis and control. Based on the knowledge obtained from the various research projects discussed in this paper, avenues for future research paths are also discussed to further tighten the bridge between theory and practice in the domain of Operations Research in general and dynamic project scheduling in particular.


TRANSITION AND REVERSION OF JAPANESE CORPORATE RATING STRUCTURE UNDER THE RECENT CREDIT CRISES

Motohiro Hagiwara, Yasuhiro Matsushitaand Katsuaki Tanaka

School of Commerce, Meiji University, Tokyo, Japan

SET Software, Nagoya, Japan

Faculty of Business Administration, Setsunan University, Osaka, Japan

Abstract. This study attempts to know the details of transition of rating structure under recent credit crises starting from BNP Paribas shock in August 2007 using Artificial Neural Network (ANN). This study checks the transition of recent bond rating structure under the credit crises based on accounting information giving Altman Z-score. Japanese corporate bond ratings are transformed to normally distributed variables using published 5-Year actual default probability, and are modelled as functions by key ratios giving Z-score. As remarkable findings, rating structure of S&P experienced serious shock in 2009 and reversion to the structure of training period in 2010. But time inconsistency of rating structures remains even in 2010. On the other hand, other agencies become more and more estranged under the crises.


EFFICIENCY EVALUATION OF GREEK COMMERCIAL BANKS USING DATA ENVELOPMENT ANALYSIS

Anastasios D. Varias and Stella Sofianopoulou

University of Piraeus, Piraeus, Greece

Abstract. The purpose of this work is to evaluate the efficiency of the biggest commercial banks that operated in Greece at the financial year 2009. The method used is Data Envelopment Analysis. Each bank is modelled as a linear system with multiple inputs and outputs. The data used was derived from the balance sheets, income statements and the annual report of each commercial bank. These data include the interest expenses, fixed assets, deposits etc. To estimate the relative efficiency of the chosen DMU’s the MS Excel add-in program xIDEA 2.1 is employed. The results indicate several inefficiencies that have no direct relation to the profitability of such institutions.


A COMPUTATIONALLY EFFICIENT MODELLING-TO-GENERATE-ALTERNATIVES METHOD USING THE FIREFLY ALGORITHM

Raha Imanirad and Julian Scott Yeomans

OMIS Area, Schulich School of Business, York University, Toronto, Canada

Abstract. In solving many practical mathematical programming applications, it is preferable to formulate numerous quantifiably good alternatives that provide very different perspectives to the problem. This is because decision-making typically involves complex problems that are riddled with incompatible and inconsistent performance objectives and possess competing design requirements which are very difficult – if not impossible – to quantify and capture at the time that the supporting decision models are constructed. There are invariably unmodelled design issues, not apparent at the time of model construction, which can greatly impact the acceptability of the model’s solutions. Consequently, it is preferable to generate several alternatives that provide multiple, disparate perspectives to the problem. These alternatives should possess near-optimal objective measures with respect to all known modelled objective(s), but be fundamentally different from each other in terms of the system structures characterized by their decision variables. This solution approach is referred to as modelling-to-generate-alternatives (MGA). This study demonstrates how the biologically-inspired, Firefly algorithm can be used to efficiently create multiple solution alternatives that both satisfy required system performance criteria and yet are maximally different in their decision spaces.


EBITDA BASED ON COMMERCIAL MARGIN PREDICTION BY HYBRID MODEL FOR READY MIXED CONCRETE BUSINESS

Pratchaya Chanprasopchai and Walailak Atthirawong

King Mongkut’s Institute of Technology Ladkrabang (KMITL), Bangkok, Thailand

Abstract. This paper proposes EBITDA calculation methodology based on commercial margin (CM) prediction by hybrid ANNs -regression and hybrid multiple regression (MR) -ANNs models for ready mixed concrete (RMC) business, which both hybrid models are suited to evaluate EBITDA. The CM accuracy performance was measured by mean absolute percentage error (MAPE), and root mean square error (RMSE), that can imply to calculate EBITDA. The CM from both models was conducted to calculate EBITDA and compared for business proposed. The EBITDA results reveal that mean absolute deviation (MAD), and tracking signal of hybrid MR -ANNs model is the lowest. As such, it can be claimed that the hybrid MR -ANNs model is more suitable approach to evaluate EBITDA based on commercial margin prediction in RMC business between two techniques.


TRAINEES ASSIGNMENT PROBLEM IN INFORMATION TECHNOLOGY (IT) SERVICE ORGANIZATIONS

Mangesh Gharote, Rahul Patil and Sachin Lodha

SJMSOM, Indian Institute of Technology-Bombay, Mumbai, India

Tata Consultancy Services Ltd, Pune, India

Abstract. This paper addresses the trainees assignment problem of an IT service firm. A Linear Programming (LP) model is developed to assign trainees to projects as per the requirements, considering their skill set and location preferences. The resulting LP model is solved using the actual cost data from the firm. This paper also discusses the implications of human resource allocation policies on the total cost.


THE DISCRETE-EVENT SIMULATION MODEL OF A HEALTH SCREENING CENTER

Thanon Wongsammacheep, Juta Pichitlamken and Waressara Weerawat

Department of Industrial Engineering, Kasetsart University, Bangkok, Thailand

Department of Industrial Engineering, Mahidol University, Nakhon Pathom, Thailand

Abstract. The health screening center is the first department that patients come into contact before going to other departments. Patients sometimes complain about long waiting times at this center. We develop a discrete-event simulation model of the health screening center to support the decision making process of the hospital management. It is designed such that it can readily be used for testing new patients’ routing plans inside the health screening center. Input data is collected from electronic records and interviews with staff. The simulation model is validated by considering the average total times in the system of one health checkup package. The two sample t-test of the empirical data and the simulation results give the p-value of 0.1; therefore, the simulation model can adequately represent the actual system.


A SIMULATION MODEL OF A HOSPITAL’S CLINICAL LABORATORY

Kanyarat Luangmul, Juta Pichitlamken and Waressara Weerawat

Department of Industrial Engineering, Kasetsart University, Bangkok, Thailand

Department of Industrial Engineering, Mahidol University, Nakhon Pathom, Thailand

Abstract. We develop a simulation model of a clinical laboratory for a complete blood count (CBC) test in a large private hospital. The model will be used for experimenting with new lab layouts and new work processes for the CBC test. The turnaround time is defined as total time in process. The average value from the simulation model is 72.37± 87.9 minute compared with 73.08 ± 22.9 minute from the empirical data. We validate our model using the 2 sample t-test yielding the p-value of 0.8.


COMPETENCE BUILDING WITH THE USE OF NURSE RE-ROSTERING

Hilde Elise Sæther Lilleby, Patrick Schittekat, Tomas Eric Nordlander, Lars Magnus Hvattum and Henrik Andersson

Department of Industrial Economics & Technology Management, NTNU, Trondheim, Norway

SINTEF ICT, Department of Applied Mathematics, Oslo, Norway

Abstract. The global nursing shortage makes efficient use of these resources vital. Good nurse rosters assist but are often static and span over a long period while the daily personnel situation is more dynamic: for example, nurses get sick or take short notice days off. Commonly, these absences are handled by hiring extra nurses when needed. However, earlier analysis has shown that nurse rotation in combination with hiring is a much more efficient solution when there exist a tool that efficiently re-rosters already planned nurses. Moreover, re-rostering gets easier if the hospital possesses the best mix of experience level and special skills. In other words, a more suitable competence profile makes re-rostering more beneficial. Nurse rotation (regularly working in another department) builds up competence, which allows for a more robust competence profile—departments become better suited to handle future personnel absences. We present a prototype that optimizes the competence profile using a stochastic model: it finds the optimal competence profile under the assumption that nurse rotation is allowed and the hospital can buy in competence. Each profile is evaluated by different absence scenarios. Our preliminary experiments show how a more robust competence profile is able to result in a 40 % cost reduction for the hospital while retaining at minimum the quality of care.


DESIGNING ROTATABLE MACHINE LAYOUT IN MULTIPLE ROW ENVIRONMENT

Srisatja Vitayasak and Pupong Pongcharoen

Centre of Operations Research & Industrial Applications, Department of Industrial Engineering, Faculty of Engineering, Naresuan University, Pitsanulok, Thailand

Abstract. Arranging of non-identical machines in the limited manufacturing shop floor is one of the essential plant designs to minimise handling distance. Shorten material handling distance can be considered as a key performance index of internal logistic activity in manufacturing industry. It leads to the efficiency of productivity and related costs. The layout design is also known as facility layout problem and classified into Non-deterministic Polynomial-time hard Problem. Most previous research related to machine layout has been focused on fixed machine orientation, which means that machine can not be rotated. The rotatable rectangular-shape machines usually have an affect on the location of pick up and drop off point, area requirement and material handling distance. This paper presents the application of Genetic Algorithm (GA) for designing rotatable rectangular machine layout in a multiple-row environment aimed to minimise the total material handling distance required for manufacturing products. Computational experiments were conducted using five datasets with different rectangular to square (R/S) ratios. The experimental results obtained were analysed. The average material handling distance decreased depending on the rotatable constraints and R/S ratio.


HEURISTIC ORDERING FOR ANT COLONY BASED TIMETABLING TOOL

Thatchai Thepphakorn and Pupong Pongcharoen

Centre of Operations Research & Industrial Applications, Department of Industrial Engineering, Faculty of Engineering, Naresuan University, Pitsanulok, Thailand

Abstract. Course timetabling is one of the core operations faced by educational institution. Timetabling problems periodically arise every academic term and are usually solved by academic staff with/without course timetabling tool. The Ant Colony based Timetabling Tool (ANCOTT) has been developed for the university course timetabling. One of ant colony optimisation variants called Rank-based Ant System (AS-rank) has been applied and embedded into the ANCOTT program to seek the feasible timetables with the lowest number of soft constraint violations. New multiple heuristic orderings between the Largest Unpermitted Period Degree first (LUPD) and the Largest Enrollment first (LE) were combined and proposed for reducing the number of infeasible timetables faced by AS-rank. Advanced statistical tools were applied to investigate the optimal setting of AS-rank’s parameters. A sequential experiment was designed and carried out using three course timetabling datasets adopted from the third track of the International Timetabling Competition (ITC2007). Serial LUPD+LE and Parallel LUPD&LE outperformed all single heuristic orderings and also produced almost 100% feasible timetables with quicker computational time than that of conventional orderings.


ADDITIVE PREEMPTIVE POSSIBILISTIC PROGRAMMING IN ASSEMBLE-TO-ORDER PRODUCTION PLANNING: A CASE STUDY

Sudatorn Sresaard and Busaba Phruksaphanrat

Industrial Statistics and Operational Research Unit (ISO-RU) in Industrial Engineering Department, Faculty of Engineering, Thammasat University, Thailand

Abstract. Additive Preemptive Possibilistic Linear Programming (APPLP) model for Assembly-to-Order (ATO) production planning is proposed in this research. Uncertainties in ATO environment make difficulties in creating a precise plan. So, most of the factories under these conditions always face problems of raw material shortage and unsatisfied demands, including the case study company. Imprecise operating costs and demands are considered in the proposed model. The model attempts to maximize profit by transforming the fuzzy objective function to three crisp objective functions, which are maximizing the opportunity of obtaining the higher profit, minimizing the risk of obtaining the lower profit and maximizing the most possible value of profit. Then, preemptive priority in the additive fuzzy programming is applied for achieving the satisfaction level of each priority that the decision maker satisfies. It makes easiness in finding a compromise solution and adjusting a satisfaction level. Alternative compromise solutions can be easily generated for the decision maker.


FUZZY MULTI-OBJECTIVE ASSIGNMENT MODEL FOR MANUFACTURING CELL FORMATION WITH DUPLICATED MACHINES

Wuttinan Nunkaew and Busaba Phruksaphanrat

Industrial Statistics and Operational Research Unit (ISO-RU) in Industrial Engineering Department, Faculty of Engineering, Thammasat University, Thailand

Abstract. This research proposes part-family and machine-cell formation in a manufacturing cell design with the consideration of duplicated machines. In existing methods, part families and machine cells are firstly determined. Then, the incidence matrix of the cell formation is reconsidered for assignment of the duplicated machines. These ways are difficult and complicated. Moreover, they not guarantee for efficient solutions. So, the effective fuzzy multi-objective assignment model is developed for solving the cell formation problem with simultaneous assignment of the duplicated machines. Two crucial performance measures of a perfect grouping called exceptional elements and void elements, are considered and applied to the proposed model. Preemptive fuzzy goal programming is also used. Significant constraints are applied for enhancing the duplicated machine assignment process. Then, the decision maker can easily solve the cell formation problem and assign the duplicated machine, simultaneously. Adjustment to find the preferred solutions can be effectively done. The numerical example is also demonstrated to show the easiness, efficiency and flexibility of the proposed model.


PLANNING OF WORK SCHEDULES FOR TOLL BOOTH COLLECTORS

Juthathip Vittawasakul and Juta Pichitlamken

Department of Industrial Engineering, Kasetsart University, Bangkok, Thailand

Abstract. We consider the workforce scheduling problem of toll booth collectors. Integer programming models are formulated to determine the number of toll booths required to satisfy vehicle demands and then the manpower shifts. Appropriate work schedules may reduce the number of employees and save labour costs but still respond to a minimum number of employees required in each period. The proposed schedule decreases the number of shifts and reduces the work hours of staff and substitutes by 70.4 (6.21%) and 8 (4.55%) hours, respectively. However, this new schedule will increase the work hours of stand-by by 46.3 hours (16.42%). Total work hours of the new schedule is 1,560.3 hours which is reduced from that of the current schedule by 32.1 hours (2.02%).


AN INVESTIGATION OF OPTIMUM CUTTING CONDITIONS IN FACE MILLING ALUMINUM 7075-T6 USING DESIGN OF EXPERIMENT

Surasit Rawangwong, Jaknarin Chatthong, R. Burapa and W. Boonchouytan

Department of Industrial Engineering, Faculty of Engineering, Rajamangala University of Technology Srivijaya, Songkhla, Thailand

Abstract. The purpose of this research was to investigate the effect of main factors of the surface roughness in aluminum 7075-T6 face milling. The results of the research could be applied in the manufacture of automotive components and mold industry. The study was conducted by using computer numerical controlled milling machine with 63 millimeter diameters fine type carbide tool with twin cutting edge. The controlled factors were the speed, feed rate and the depth of cut which the bite cutter was not over 1 mm. For this experiment we used factorial designs and the result showed that the factors effecting the surface roughness were the feed ratio and the speed while the depth did not effect the surface roughness. Furthermore, the result of the test showed that the surface roughness was likely to reduce when the speed was 2,930 rpm and the feed rates was 808 mm/min. The result of the research led to the linear equation measurement value which was Ra = 0.123 -0.000039 Speed + 0.000108 Feed. The equation formula should be used with the speed in the range of 2,021-2,930 rpm, feed rate in the range of 808-1,172 mm/min and the depth of not over 1 mm. When the equation was used to confirm the research results, it was found that the mean absolute percentage error (MAPE) of the surface roughness obtained from the prediction comparing to the value from the experiment was 4.03 percent, which was less than the specified error and it was acceptable.


COORDINATION BUYER AND SUPPLIER WITH VENDOR MANAGED INVENTORY FROM THE NET PRESENT VALUE PERSPECTIVE

Chaowalit Hamontree, Djamila Ouelhadjand Patrick Beullens

Logistics and Management Mathematics Group, Department of Mathematics, University of Portsmouth, Lion Terrace, Portsmouth, UK

School of Mathematics/School of Management, University of Southampton, UK

Abstract. This paper considers a two-level supply chain consisting of single-buyer single-supplier under deterministic conditions and constant demand rate. The main objective is to find optimal lot-sizing decisions and inventory policies, which derive from Net Present Value (NPV) framework. We study the models in two cases; individual and coordination decisions. In the individual case the buyer and the supplier act independently to optimise their profits or costs; while in the coordination case both actors are considered to determine the optimal variables to achieve profit (or cost) functions. The classic single-buyer single-supplier inventory model is applied in the individual model and Vendor-Managed Inventory (VMI) strategies are applied in the coordination models where the supplier is authorized to manage the buyer’s inventory and make all the replenishment decisions. The experimental results have shown that the optimal solution is given from a coordinated VMI approach derived from NPV framework. Furthermore, the numerical examples have shown that the optimal solution of VMI+ achieves a very good coordination. This model guarantees that the profit function for both players remains the same without coordination and increases as a result of coordination.


INCREASING UNCERTAINTY IN CASH FLOW SIMULATION-BASED VOLATILITY ESTIMATION FOR REAL OPTIONS: ACTUAL INCREASE IN VOLATILITY OR SYMPTOM OF EXCESS UNRESOLVED AMBIGUITY UNCERTAINTY

Tero Haahtela

Aalto University, BIT Research Centre, Helsinki, Finland

Abstract. This paper investigates how cash flow simulation-based volatility estimation methods work in the case of decreasing or increasing volatility during a project. Whereas contingent claims analysis-based methods work correctly in both situations, this does not necessarily hold for cash flow simulation-based methods, because the underlying asset value may not be known accurately in the beginning. Misinterpreting ambiguity for volatility may cause decision makers to significantly overestimate volatility, and therefore overestimating a project value with options. Ambiguity may also cause erroneous investment outlays as the decision maker cannot follow precisely the underlying asset value, because it is not an observable market price. This is problematic especially in the case of increasing volatility. This paper describes the dilemma and suggests some managerial advice on how to detect, understand, and mitigate the problem.


STOCHASTIC SEARCH MECHANISMS ON THE BEE ALGORITHM FOR OPTIMISING NOISY MULTI-RESPONSE SURFACES

Pasura Aungkulanon and Pongchanun Luangpaiboon

Industrial Statistics and Operational Research Unit (ISO-RU), Department of Industrial Engineering, Faculty of Engineering, Thammasat University, Thailand

Abstract. In this paper, bee algorithm (BA) and its hybridisations has been used for searching the near optimum in uni-and multi-response surfaces with the noise on the responses. Cooperation and communication of the members of honey bees in food foraging are similar to the optimisation process which results in finding a global solution as determined by an objective function. So far the BA is typically applied to continuous process optimisation problems. In order to investigate the performance of BA on complex and multi-objective optimisation problems, an attempt is made in this paper. In this paper a brief review of BA is first given and the development of the BA for solving noisy multi-response surfaces is then presented via the desirability function approach. Two structural mechanisms from other metaheuristics of the firefly (FA) and variable neighbourhood search (VNS) via the vertical transportation strategy were then proposed for its solution searching improvement. A computational study is carried out and the results are compared with several performance measures to demonstrate the effectiveness and robustness of the hybrid algorithms. The results indicate that the hybrid with the FA is a powerful search and optimisation technique. It yields better solutions when compared to those obtained by original algorithms.


AN EVALUATION OF HOSPITAL LOCATIONS WITH RESPECT TO PROXIMITY, COVERAGE, AND EQUALITY OF SERVICE

H. A. Eiselt, Joy Bhadury and Mark L. Burkey

Faculty of Business Administration, University of New Brunswick, Fredericton, NB, Canada

Bryan School of Bus & Econ, University of North Carolina -Greensboro, Greensboro, NC, USA

School of Business and Economics, North Carolina A & T State University, Greensboro, USA

Abstract. This article evaluates the present locations of general hospitals in four U.S. States. The actual locations of hospitals are compared to the optimal location pattern, if all hospitals would be located today. The basis of our comparison includes accessibility, proximity, and "equity" criteria. It turns out that while hospitals were located over time and based on often political considerations, their present locations are relatively close to optimality. The paper concludes with some thoughts on potential extensions of our work, including hospital closures, expansions, and others.


DIFFERENCES BETWEEN FINANCIAL OPTIONS AND REAL OPTIONS

Tero Haahtela

Aalto University, BIT Research Centre, Helsinki, Finland

Abstract. Real option valuation is often presented to be analogous with financial options valuation. The majority of research on real options, especially classic papers, are closely connected to financial option valuation. They share most of the same assumption about contingent claims analysis and apply close form solutions for partial difference equations. However, many real-world investments have several qualities that make use of the classical approach difficult. This paper presents many of the differences that exist between the financial and real options. Whereas some of the differences are theoretical and academic by nature, some are significant from a practical perspective. As a result of these differences, the present paper suggests that numerical methods and models based on calculus and simulation may be more intuitive and robust methods with looser assumptions for practical valuation. New methods and approaches are still required if the real option valuation is to gain popularity outside academia among practitioners and decision makers.


COMPARISON OF EFFECTIVENESS OF DIFFERENT IMPLEMENTATIONS OF A HEURISTIC FOREST HARVEST SCHEDULING SEARCH PROCEDURE WITH DIFFERENT NUMBER OF DECISION CHOICES SIMULTANEOUSLY CHANGED PER MOVE

Jordi Garcia-Gonzalo, José Guilherme Borges, Wesley Hilebrand and João Palma

Instituto Superior de Agronomia, Centro de Estudos Florestais, Universidade Técnica de Lisboa, Lisboa, Portugal

Abstract. This study presents a comparison of the performance of a heuristic technique (Simulated Annealing) with one to three decision choices simultaneously changed in each move. This means that the heuristic was implemented to look for better solutions by changing the treatment schedule in up to three stands in each move. For testing purposes these different implementations where applied to a real harvest scheduling problem. The test forest consists of 1000 eucalypt stand stands extending over 11873 ha in Portugal. The test problems encompassed a 30 years temporal horizon. The management planning objectives encompassed the maximization of soil expectation value as well as the regulation of harvest flows. Several starting parameters were tested (e.g. initial temperature, cooling schedule, penalties and stopping criteria) in over y computer runs. Preliminary results showed that a one-stand move (i.e. changing the treatment schedule in one compartment at each move) was better than two-stand moves in any cooling schedule. In addition faster cooling schedules where better than slower cooling schedules.


EFFICIENT MULTIVARIATE MODELING OF CROSS BORDER EFFECTS IN EUROPEAN BOND VOLATILITY SPILLOVER: A MULTIPLE OBJECTIVE ARTIFICIAL NEURAL NETWORK APPROACH

Gordon H. Dash, Jr. and Nina Kajiji

Finance and Decision Sciences Area, College of Business Administration, University of Rhode Island, Kingston, RI, USA

The NKD-Group, Inc. and Department of Computer Science and Statistics, University of Rhode Island, Kingston, RI, USA

Abstract. In this paper we extend prior efforts to engineer an efficient mapping of volatility transmission across various western-and central-European government bond markets. The univariate Bayesian-enhanced multiple-objective K4-RANN has been a standard to produce an efficient minimization of the ill-effects of multicollinearity while attaining maximum smoothness in nonparametric time series analysis. This research introduces a multivariate extension to the K4 algorithm; an extension which permits multiple target variables to be specified in the estimating equation. The new K7-MRANN is employed to re-examine bond volatility spillover effects previously obtained from univariate parametric-and artificial neural network based conditional volatility investigations. A prior K4-RANN estimation produced residuals that were nearly devoid of latent economic effects along with network weights that both corroborated and extended prior parametric findings of a weak US spillover effect into established European bond markets. The K7-MRANN findings presented in this research report model residuals that are clearly linearly independent and devoid of any latent interpretation. The signed model weights produced by a simultaneous multivariate RANN provide convincing evidence of the uniform effect of a negative U.S. bond market spillover into the aggregate Europe bond markets as well into specific Euro-domestic bond markets. The multivariate modelling efforts also provide a measurable view into the intra-Europe sovereign spillover effects.


ON THE BEHAVIORAL SPECIFICATION AND MULTIVARIATE NEURAL NETWORK ESTIMATION OF COGNITIVE SCALE ECONOMIES

Nina Kajiji and Gordon H. Dash, Jr.

The NKD-Group, Inc. and Department of Computer Science and Statistics, University of Rhode Island, Kingston, RI, USA

Finance and Decision Sciences Area, College of Business Administration, University of Rhode Island, Kingston, RI, USA

Abstract. It is universally accepted that sovereign economic development is inextricably linked to efficient education production. In the United States, the task of designing policies to enhance education attainment by students is an effort that is largely reliant upon state-wide and local policy initiatives. Because policy-making skills vary across these disaggregated educational bodies, it has become increasingly important for government leaders at all levels to assist regional and state leaders in the process of identifying effective policy metrics. This paper presents evidence that a multivariate artificial intelligence-based model produces more plausible estimates of traditional production elasticity metrics when compared to its parametric alternative. The scale elasticity metrics obtained by solving a multivariate OLS (MOLS) model are compared directly to those generated by solving a multivariate nonparametric multiple objective radial basis function artificial neural network model (MRANN).


A HYBRID METAHEURISTIC FOR THE PARTITIONING PROBLEM WITH HOMOGENEITY CONSTRAINTS ON THE NUMBER OF OBJECTS

María Beatriz Bernábe Loranca, David Pinto Avendaño, Elias Olivares, Benitez, Rogelio González Velázquez, J. L. M. Flores and J. R. Vanoye

Facultad de Ciencias de la Computación, Benemérita Universidad Autónoma de Puebla, Puebla, México

Posgrado de Logística y Dirección de la cadena de Suministro, Universidad Popular Autónoma del Estado de Puebla, Puebla, México

Abstract. Partitioning is a combinatorial optimization problem that has been widely discussed because of two main aspects: its computational complexity and its application to diverse kind of problems. In addition, partitioning has been used on the solution of some clustering issues that arise when dealing with territorial design problems. Besides the particular properties of the classic partitioning, in some cases, it is needed to consider additional constraints that need to be solved, such as the homogeneity of the number of objects that made up the final clusters. In this paper, we present a solution applied to partitioning problems having the moderate balance constraint, by determining thresholds which allows it obtaining a well defined range on the dimensionality of the groups. The aforementioned problem with constraints is considered to have a high computational complexity and, therefore, in order to obtain approximate solutions for this problem, we have incorporated a particular hybrid metaheuristic of partitioning that combines simulated annealing and variable neighborhood search. We present the problem, a solution proposal, the mathematical model and a set of experimental tests for samples taken from a study case based on 469 geographical data of the Metropolitan Zone of the Toluca Valley.


A CONTRIBUTION TO GENERAL REPAIR MODELS AND THEIR APPLICATION

Frank Beichelt

School of Statistics and Actuarial Science University of the Witwatersrand, Johannesburg, Republic of South Africa

Abstract. Most maintenance policies use maintenance actions, which only comprise the extreme maintenance measures replacements and minimal repair. In this paper, we model maintenance actions, which are ‘in between’. This will be done by a virtual age method derived from the failure rate, and by a geometric time scaling method. These methods are applied to the analysis of cost-efficient maintenance policies.


INTEGRATED LOCATION DISTRIBUTION AND INVESTMENT MODEL FOR NEW DISTRIBUTION CENTERS USING POSSIBILISTIC PROGRAMMING

Chakkapat Chuanin and Busaba Phruksaphanrat

Industrial Statistics and Operational Research Unit (ISO-RU) in Industrial Engineering Department, Faculty of Engineering, Thammasat University, Klongluang, Thailand

Abstract. Distribution Center (DC) plays a key role in reducing logistics cost for an organization. Moreover, it also can be used to increase competency in customer service. Normally, location, distribution and investment problems are considered separately. However, construction a new DC, not only investment cost but also location and distribution plans of DCs are necessary to be considered simultaneously in order to increase efficiency of a logistics system. So, in this research, an integrated location distribution and investment model for new DCs using multi-objective possibilistic mixed integer programming is proposed. Three objectives are considered; minimizing net present value of the total cost, maximizing customer service based on supplied quantities and maximizing customer service level based on ability to response customers. These are considered under uncertain cost coefficients. Possibilistic programming is used to deal with these uncertainties and ε-constraint method is applied to find a compromise solution. Decision maker can select the appropriate levels of service based on supplied quantities and the ability to response customers. Then, the suitable investment plan for each year, location decision and distribution plan can be generated concurrently. A case study is also illustrated to show the effectiveness on the proposed model.


FUZZY FACILITY LOCATION-ALLOCATION FOR BIOMASS TRANSFORMATION PLANT OF THE CENTRAL REGION OF THAILAND

Pornchai Hongwattanakul and Busaba Phruksaphanrat

Industrial Statistics and Operational Research Unit (ISO-RU) in Industrial Engineering Department, Faculty of Engineering, Thammasat University, Klongluang, Thailand

Abstract. Thailand is the country which has agriculture all over the country. Agricultural residues can be used as biomass energy which can help Thailand to reduce cost of importing energy. So, government is trying to encourage in investment of biomass transformation plants and use of clean energy. Biomass pellet production is the one of selected technologies for promotion in Thailand. In generating the appropriate promotion for investors, government needs to know where the transformation plants should be located. These places can be able to reduce existing logistics cost of biomass energy used. So, the mixed integer programming model for plant location-allocation for biomass transformation factories in the central of Thailand is proposed in this research. Both sources of supply and demand from industrial parks are concerned. Straw is considered to be transformed to pellet due to its high quantity energy density. The proposed model can find the appropriate location-allocation and size of the plants. Moreover, uncertainty of cost coefficients are also determined using the possibilistic distributions. Three types of solution; pessimistic, most-likely optimistic are proposed and discussed.


PLANNING ELECTIVE SURGERIES IN A PORTUGUESE HOSPITAL: STUDY OF DIFFERENT MUTATION RULES FOR A GENETIC HEURISTIC

Inês Marques, Maria Eugénia Captivo and Margarida Vaz Pato

Centro de Investigação Operacional, Universidade de Lisboa, Lisbon, Portugal

FCTS/FEG, Universidade Lusófona de Humanidades e Tecnologias, Lisbon, Portugal

Faculdade de Ciências, Universidade de Lisboa, Lisbon, Portugal

Instituto Superior de Economia e Gestão, Universidade Técnica de Lisboa, Lisbon, Portugal

Abstract. Reduced budgets in the healthcare sector pressure health institutions to an efficient use of resources. The operating theatre is a hospital unit that represents a great proportion of the hospital budget. Furthermore, it is a central service with connections and implications in the service of many other hospital units. A more efficient use of the operating room becomes of great relevance within the hospital administration. This work develops a genetic algorithm to address the elective surgery planning problem arising in a public hospital in Lisbon. Different mutation rules are discussed. Some results obtained with the hospital’s data are presented.


OPTIMISATION OF A STRICTLY CONCAVE QUADRATIC FORM

Belabbaci Amel, Djebbar Bachirand Mokhtari Abdelkader

University of Laghouat. Laghouat, Algeria

University of sciences and technology, Oran, Algeria

Abstract. In this study, we give a method which allows finding the exact optimal solution of a strictly concave quadratic program. The optimization of a strictly concave program is based on the localization of the critical point. If the critical point doesn't belong to the feasible solution set the projection onto the hyperplanes passing through the nearest vertex to the critical point gives exactly the optimal solution.


SIMPLE FUZZY INPUT SCORECARD FOR INTELLECTUAL PROPERTY RIGHTS EVALUATION

Mikael Collan

School of Business, Lappeenranta University of Technology, Lappeenranta, Finland

Abstract. Scorecards are an often used, structured, reporting method for the analysis and assessment of performance. In this paper we show how an advanced, but simple IPR scorecard that uses fuzzy numbers as inputs can be built and used in the analysis, evaluation, and the management of IPR resources. The scorecard has an intuitively understandable graphical presentation of the scorecard information and the final score -a triangular fuzzy number that is a distribution able to show estimation imprecision. Meaningful and intuitively understandable single number information can be extracted from the resulting distribution. As a result the scorecard offers enhanced decision support for the IPR manager for IPR resources that may be very hard to analyze in light of vague cash-flow information.


UNIVERSITY EFFICIENCY EVALUATION WITH USING ITS REPUTATIONAL COMPONENT

Irina Abankina, Fuad Aleskerov, Veronika Belousova, A. B. Osmolovskaya, V. Petruschenko, D. Ogorodniychuk, V. Yakuba and K. Zinkovsky

National Research University (NRU HSE), Moscow, Russian Federation

Institute of Control Science Russian Academy of Science, Moscow, Russian Federation

Abstract. We estimate efficiency scores for Russian universities based on data set of input and output criteria by using Data Envelopment Analysis. In addition, we use a reputation index as another indicator of a university’s productivity. To construct it, 4000 contexts are analyzed and 13 reputation criteria are found. The threshold procedure is used to aggregate them into a reputation indicator. Factors which lead a university to be efficient are studied.


COMMON-KNOWLEDGE, COMMUNICATION AND MORAL HAZARD IN PRINCIPAL-AGENT MODEL

Takashi Matsuhisa

Department of Natural Science, Ibaraki National College of Technology, Ibaraki, Japan

Abstract. This article re-examines a principal-agent model with moral hazard from the epistemic point of view. It highlights hidden conditions for a possible resolution of the moral hazard between the principal and the agents. We show that the moral hazard in the principal agent model under uncertainty will not be appeared if the principal and agents could share fully information on their expected marginal costs in the following two cases: first they commonly known the marginal expected costs and secondly they communicate the costs as long run.


A BRANCH AND BOUND APPROACH FOR THE SEQUENTIAL ORDERING PROBLEM

Marco Mojana, Roberto Montemanni, Gianni Di Caro and Luca M. Gambardella

Dalle Molle Institute for Artificial Intelligence (IDSIA), Manno, Switzerland Università della Svizzera Italiana, Lugano, Switzerland

Abstract. Many different real problems can be modeled in mathematical terms as a Sequential Ordering Problem. This combinatorial optimization problem can be seen as a scheduling problem with precedence constraints among the jobs. In the present paper a branch and bound method that exploits the structure of the precedence constraints is introduced and tested on some benchmark instances commonly adopted in the literature.


A HYBRID VARIABLE NEIGHBORHOOD SEARCH PATH-RELINKING FOR SOLVING THE CAPACITATED SINGLE ALLOCATION HUB LOCATION PROBLEM

Ornurai Sangsawang and Sunarin Chanta

Department of Industrial Management, King Mongkut’s University of Technology North Bangkok, Thailand

Abstract. In this paper, the hybrid VNS-Path Relinking algorithm is proposed for solving the capacitated single allocation hub location problem (CSAHLP). The CSAHLP is related with locating a set of capacitated hubs and allocating spoke nodes to the hubs. In this problem, the number of locations is not known priori, and the amount of flow collected in the hub is restricted. The objective is to minimize total transportation costs by considering the fixed costs of establishing hubs. An effective Hybrid VNS-Path Relinking algorithm is investigated on AP data set with respect to solutions.


SHELTER-SITE SELECTION DURING FLOOD DISASTER

Sunarin Chanta and Ornurai Sangsawang

Department of Industrial Management, King Mongkut’s University of Technology North Bangkok, Thailand

Abstract. During flood disaster, people in affected areas seek a safe place to stay. Most people evacuate to shelter, which is a public safe place usually organized and provided by government. Not only provides a place for evacuees, but shelter also works as a distributer to provide necessary supplies for people staying in their houses in surrounding effected area. In this paper, we propose an optimization model to find appropriate locations of temporary shelters. The objectives are to maximize the number of flood victims that can be covered or can reach a shelter within a fixed distance and to minimize the total distance of all flood victims to their closest shelters. Since the problem is formulated as a bi-objective programming model, the epsilon-constraint is selected to solve the problem. We present a real case study obtained data from severe flooding in Thailand, 2011.


OPTIMIZED PATIENT TRANSPORT

Patrick Schittekat and Tomas Eric Nordlander

SINTEF ICT, Department of Applied Mathematics, Oslo, Norway

Abstract. In this article, we present an optimized patient transport problem found in most large hospitals. The proposed solution aims to reduce patient waiting time and transporter idle time resulting in a higher throughput of patient transports and better patient care.


ADVANCES IN STOCHASTIC MODELLING FOR MAINTENANCE

Frank Beichelt

University of the Witwatersrand, Johannesburg, Republic of South Africa

Abstract. The development of stochastic models for the cost-efficient maintenance of technical systems started in the fifties of the past century, quite in line with some other classes of mathematical models in operations research. Since then, the number of papers published on the subject has been growing exponentially year by year. This trend does not imply that there is a real practical demand for most of these papers. The models analyzed have become more and more artificial, and the majority of them have no real chance of ever being applied, since they lack a genuine connection to the industrial reality. The reason for this situation is that maintenance policies–just as any other mathematical models of operations research– need to be tailored to the specific practical situation. For that purpose, however, general but simply structured theoretical model classes in maintenance theory have proved to be very helpful. They provide the general framework for optimally organizing maintenance measures, both with regard to cost, reliability, availability, useful life and other criteria. In this talk, three general approaches to maintenance will be discussed in more detail. The first one is the concept of a general repair. Contrary to most of the maintenance policies proposed so far, it takes into account that repairs do not completely renew a system and that they have a non-negligible influence on its future failure behaviour. The second approach is called total or cumulative maintenance cost limit replacement. For making the decision whether to repair or to scrap and replace a system, only its maintenance cost history is taken into account, i.e. failure, maintenance, and other ‘loss’ times and the character of maintenance actions are only relevant with regard to their cost implications. The third model class takes into account system warranty provided by the manufacturer for scheduling preventive maintenance.


PARALLEL TABU SEARCH FOR THE CAPACITATED VEHICLE ROUTING PROBLEM

Jianyong Jin, Teodor Crainicand Arne Løkketangen

Molde University College, Specialized University of Logistics, Molde, Norway

CIRRELT, UQAM, Montreal, Canada

Abstract. The Capacitated Vehicle Routing problem, CVRP, is one of the basic problems in transportation optimization. We present a guided cooperative parallel tabu search algorithm for solving this problem. Diversification and intensification mechanisms are applied to the single search threads, and guidance is provided by knowledge extracted from previously found good solutions. Our solution methods obtain competitive or better solutions to a set of benchmark problems from literature. Detailed computational results will be presented at the conference.


© ORLab Analytics Inc