ICAOR’15 ABSTRACTS

7TH INTERNATIONAL CONFERENCE ON APPLIED OPERATIONAL RESEARCH

15-17 JULY 2015, VIENNA, AUSTRIA


IMPLEMENTING MORE SUSTAINABLE TRANSPORT FOR HOME SERVICE PROVIDERS - CHALLENGES, METHODS AND IMPLICATIONS

Christian Fikar 1 and Patrick Hirsch 1

1 Institute of Production and Logistics, University of Natural Resources and Life Sciences, Vienna, Feistmantelstr. 4, A 1180 Wien, Austria

{christian.fikar, patrick.hirsch}@boku.ac.at

Abstract. Urbanisation causes various challenges (e.g., congestions, limited parking space) for service providers performing tasks at customer's premises. Up to now, the staff uses mainly individual cars. In this talk, we investigate the optimisation of different alternative transportation concepts for a major Austrian home health care provider. These include trip sharing with walking-routes as well as routing and scheduling with bikes and public transport. The concepts lead to interesting optimisation problems due to synchronisation constraints, interdependencies between different routes and time-dependencies. The results show substantial potential to reduce the number of required vehicles, save travel time and to enable more sustainable operations.


MATHEURISTIC OPTIMIZATION FOR ROBUST HOME HEALTH CARE SERVICES

Thi Viet Ly Nguyen 1, Nihat Engin Toklu 1 and Roberto Montemanni 1

1 Dalle Molle Institute for Artificial Intelligence (IDSIA – USI/SUPSI), Galleria 2, 6928 Manno, Switzerland

{vietly, engin, roberto}@idsia.ch

Abstract. In modern societies, home health care services are becoming increasingly important. Having optimized solutions on operational issues can play a potential role in offering patients high quality medical service as well as taking in better regard, the needs of care providers. An important issue to consider is the uncertainty on the problem data. In more details, an optimal solution that was obtained under the assumption that the collected problem data are accurate, can turn out to be infeasible when it is implemented in the reality, because the data encountered in the reality might differ from the assumed data in the optimization model. In this paper, we consider the uncertainty on the availability of the nurses (e.g. they might call sick on short notice), and we follow a robust optimization approach to handle the case when nurses are unexpectedly unable to operate. A matheuristic method based on a constructive heuristic combined with a genetic algorithm and mathematical programming is proposed to provide a near-optimal solution both in terms of nurse-patient assignment and nurse scheduling and routing.


A CROWD SIMULATION FOR A LARGE-SCALE INDOOR VENUE - A CASE STUDY OF TAIPEI ARENA

Yun-Zhu Lin 1, Ming-Gin Wu 1 and Che-Fu Hsueh 2

1 Department of Civil Engineering, Chung Yuan Christian University, Taoyuan City, Taiwan

mschang@cycu.edu.tw

2 Department of Marketing and Distribution Management, Chien Hsin University of Science and Technology, Taoyuan City, Taiwan

cfhsueh@uch.edu.tw

Abstract. Previous research into crowd evacuation simulation has focused on exploring general buildings and urban pedestrian traffic. Crowd evacuation simulation of large-scale venues has been limited. This study verifies the applicability of adopting Pathfinder 2011, agent-based simulation software, to develop a simulation model for crowd evacuation at a large-scale venue. Taipei Arena was chosen for the case study. Different evacuation simulation conditions and the results demonstrate that the steering movement mode of Pathfinder is suitable for simulating the movement of heterogeneous crowds and L-shaped movement in stairways and raised seating areas. Two functions of Pathfinder, “occupant grouping” and “designation of evacuation path” can construct an occupant movement model of a large-scaled indoor venue and each evacuation path to consist of one final destination (main entrance) and one mid-point (exits in the seating area or corridor platform). Such guidance is much more important for people seated in raised seating areas on higher floors within a large-scale indoor venue than for other areas in this venue, due to the limitations of stair width and inclination.


USING MULTI-START BIASED RANDOMIZATION OF HEURISTICS TO SOLVE THE VEHICLE ROUTING PROBLEM WITH CLUSTERED BACKHAULS

Javier Belloso 1, Angel A. Juan 2, Javier Faulin 1 and Adrián Serrano 1

1 Public University of Navarra, 31006 Pamplona, Spain

{javier.belloso, javier.faulin, adrian.serrano}@unavarra.es

2 Open University of Catalonia, 08018 Barcelona, Spain

ajuanp@uoc.edu

Abstract. We consider the Vehicle Routing Problem with Backhauls (VRPB), where delivery and pick-up customers are to be served from a central depot. In particular, the group or cluster of delivery customers has to be served before the first pickup customer can be visited. Thus, the problem belongs to a sub-class called VRP with Clustered Backhauls (VRPCB). Our resolution procedure uses a multi-start approach designed to avoid the local minima and to be easily parallelizable. The algorithm employs a biased-randomized version of the classical savings heuristic, together with some local search processes. During the solution-construction process, the edges that connect one delivery customer with a pick-up customer are penalized to be chosen at a later stage. The savings list of edges is randomized using a skewed probability distribution. Some classical benchmark instances for the VRPB were selected in order to compare the efficiency of our approach.


OPTIMIZING NON-COMBATANT EVACUATION OPERATION TRANSPORTATION LOGISTICS

Bohdan L. Kaluzny 1 and Jean-Denis Caron 1

1 Defence Research and Development Canada, 101 Colonel By Dr, Ottawa, ON, K1A 0K2 Canada

Bohdan.Kaluzny@forces.ca; Jean-Denis.Caron@drdc-rddc.gc.ca

Abstract. A non-combatant evacuation operation (NEO) is designed to deploy forces in a short notice situation as a result of a deteriorating situation in an affected nation that is threatening the safety of citizens or eligible persons living abroad. An optimization model was developed to study the feasibility and logistical complexities of transportation asset usage of a NEO, enabling decision makers to better plan and prepare NEO execution.


SIGNALING INFORMATION IN A VERTICAL DECENTRALIZED SUPPLY CHAIN

Huajiang Luo 1, Weixin Shang 1 and Tian Li 2

1 Faculty of Business, Lingnan University, Tuen Mun, Hong Kong

{huajiangluo, wshang}@ln.edu.hk

2 School of Business, East China University of Science and Technology, Shanghai, China

litian@ecust.edu.cn

Abstract. We consider a decentralized supply chain with one manufacturer and one retailer who are engaged in a sequential-move game. The manufacturer and the retailer decide the wholesale price and retail margin to maximize their respective profits. The manufacturer possesses some private information about future demand, and hence enjoys advantage from both Stackelberg leadership and information over the retailer. Though the retailer cannot observe the manufacturer's private information, he can inferred it from leader's wholesale price. However, to induce the retailer to reduce retail margin, the manufacturer has to lower down the wholesale price to signal a low demand which benefits the retailer. The general research question is whether the manufacturer's profit is still higher than the retailer's profit, as in the sequential-move game without private demand information. We find that information leakage cost may outweigh the benefit from Stackelberg leadership, and characterize conditions under which the retailer's profit is higher than that of the manufacturer.


SCHEDULING GROUND-HANDLING SERVICES: A BI-OBJECTIVE APPROACH

Daniel Guimarans 1, Silvia Padrón 2, Juan José Ramos 2, Salma Fitouri-Trabelsi 3 and Liana Napalkova 2

1 Optimisation Research Group, National ICT Australia (NICTA), 13 Garden Street, Eveleigh NSW 2015, Australia

daniel.guimarans@nicta.com.au

2 Dpt. Telecommunication and Systems Engineering, Universitat Autònoma de Barcelona, Emprius 2, 08202 Sabadell, Barcelona, Spain

{silvia.padron, juanjose.ramos, liana.napalkova}@uab.cat

3 Mathématiques Apliquées, Informatique et Automatique pour l’Aérien & Air Transportation Department, L’École Nationale de l’Aviation Civile (ENAC), 7 Avenue Édouard Belin, 31055 Toulouse, France

salma.fitouri-trabelsi@enac.fr

Abstract. Preparing an aircraft for its next flight requires a set of interrelated services involving different types of vehicles. Planning decisions concerning each resource affect the scheduling of the other activities and the performance of the other resources. Considering the different operations and vehicles instead of scheduling each resource in isolation allows integrating decisions and contributing to the optimisation of the overall ground-handling process. This goal is defined through two objectives: (i) minimising the waiting time before an operation starts and the total reduction of corresponding time windows, and (ii) minimising the total completion time of turnarounds. We combine different technologies and techniques to solve the problem efficiently. A new method to address this bi-objective optimisation problem is also proposed. The approach has been tested using real data from a major Spanish airport, obtaining different solutions that represent a trade-off between both objectives. Experimental results permit inferring interesting criteria on how to optimise each resource, considering the effect on other operations. This outcome leads to more robust global solutions and to savings in resources utilization.


PUSHBACK DELAYS ON THE ROUTING AND SCHEDULING PROBLEM OF AIRCRAFT

Christofas Stergianos 1, Jason Atkin 1, Patrick Schittekat 2, Tomas Eric Nordlander 2, Chris Gerada 1 and Herve Morvan 1

1 Institute for Aerospace Technology, The University of Nottingham, Nottingham, UK

{Christofas.Stergianos, Jason.Atkin, Chris.Gerada, Herve.Morvan}@nottingham.ac.uk

2 SINTEF ICT, Department of Applied Mathematics, Oslo, Norway

{Patrick.Schittekat, Tomas.Nordlander}@sintef.no

Abstract. With the constant increase in air traffic, airports are facing capacity problems. Optimisation methods for specific airport processes are starting to be increasingly utilised by many large airports. However many processes happen in parallel and a more complex optimisation model is required, which can consider multiple processes simultaneously. This paper focuses on the importance of the pushback process in the routing process. It investigates whether taking the pushback process into consideration can predict delays that otherwise would pass unnoticed. Having an accurate model for the pushback process is important for this and identifying all of the delays that may occur can lead to more accurate and realistic models that can be used in the decision making process for ground movement operations. After testing a model with a more detailed pushback process we found that a lot of the delays are not predicted if the process is not explicitly modelled. Having a more precise model with accurate movements of aircraft is highly important for an integrated model and will allow ground movement models to be used for more reliable integrated decision making systems on airports. Minimising these delays can help airports increase their capacity and become more environmentally friendly.


INCREASING TRUST IN OPTIMIZATION BASED ATM SYSTEMS THROUGH TRAINING

Tomas Eric Nordlander 1, Amela Karahasanović 2 and Patrick Schittekat 1

1 SINTEF ICT, Department of Applied Mathematics, Oslo, Norway

{Tomas.Nordlander, Patrick.Schittekat}@sintef.no

2 SINTEF ICT, Department of Networked Systems and Services, Oslo, Norway

Amela.Karahasanovic@sintef.no

Abstract. Air Traffic Management is the complex task of safely managing the flow of aircrafts. This task needs to be solved efficiently to avoid hindering the growth of the aviation sector. Optimization-based decision support tools could assist but, this being a safety-critical and conservative domain, a high level of trust needs to be in place. The amount of trust depends on the level of automation and familiarity with the tools. We argue that this needed trust could best be built if optimization based tools are used in the training of air traffic controllers. We discuss how this training can help air traffic controllers and the model of optimization.


PRUNING RULES FOR OPTIMAL RUNWAY SEQUENCING WITH AIRLINE PREFERENCES

Geert De Maere 1 and Jason A.D. Atkin 1

1 Automated Scheduling, Optimisation and Planning Group (ASAP), School of Computer Science, The University of Nottingham, Nottingham, UK

{Geert.DeMaere, Jason.Atkin}@nottingham.ac.uk

Abstract. This paper presents a new pruned dynamic programming approach to solve real world runway sequencing problems in which delay cost differentiation between the different aircraft is applied. The pruning rules are based on more generic dominance rules and properties of the objective function that enable to significantly reduce the size of the state space. Optimal results for publically available benchmark instances are reported, and obtained at an extremely low computational cost.


PERFORMANCE OPTIMIZATION IN RETAIL BUSINESS USING REAL-TIME PREDICTIVE ANALYTICS

Nizar Zaarour 1 and Emanuel Melachrinoudis 1

1 Northeastern University, 360 Huntington Avenue, Boston, MA 02115 USA

n.zaarour@neu.edu; emelas@coe.neu.edu

Abstract. Predictive analytics has become more than a necessity in today’s competitive market and technology based world. Traditional data analysis tools such as regression analysis, numerical taxonomy, cluster analysis, and other multivariate statistical methods are not sufficient anymore to handle all the complicated intricacies and big data sets. To compensate for these shortcomings, more advanced tools have emerged, which are based on artificial intelligence, pattern recognition, statistics, machine learning, and many other data mining tools. In this paper, we develop a model that focuses on collecting real-time data for the purpose of recognizing patterns in the retail and point-of-service business, and improving the sales and customer satisfaction. We were able to identify a large number of predictors affecting the behavior of sales; however, this particular paper focuses on one specific independent variable, the number of people involved in each transaction in a restaurant business. Breakdown of the model along with the results is presented, and a summary of planned future research, involving the impact of other predictors, is discussed at the end of the paper. One of the major benefits of this research, is the win-win-win approach to make sure that we optimize the satisfaction of all parties involved: management, staff, and customers.


FRAUDULENT URL CLASSIFICATION WITH AN RBFNN

Dirk Snyman 1, Tiny du Toit 1 and Hennie Kruger 1

1 North-West University, Potchefstroom Campus, Potchefstroom, South Africa

{dirk.snyman, tiny.dutoit, hennie.kruger}@nwu.ac.za

Abstract. Phishing attacks are a social engineering scam which aim to steal sensitive information like personal, financial and social details, from unsuspecting consumers. It is based on the premise that a significant number of consumers are ignorant of the technical security practices in a digital environment. These attacks establish trust from the user in order to lull them into a false sense of familiarity in which they are more likely to freely supply their identifying information. In this study an intelligent approach to phishing detection is presented. The method is based on the information contained in the Uniform Resource Locator (URL) that points to a phishing website. These URLs are analysed and classified as malicious or safe, by a new automated Radial Basis Function Neural Network (RBFNN) construction algorithm. This technique determines the best neural network architecture for a specific classification task and uses an in-sample model selection criterion. Example URLs which were collected from real world repositories, Open Directory and Phishtank, are used to train and evaluate the neural network. The results of a cross-validation experiment setup are presented. It was found that the RBFNN algorithm outperforms a naïve Bayes classifier baseline which is considered to be a standard text classification baseline in literature due to its simple structure and linear execution.


ROBUSTNESS AND MATHEURISTICS

Roberto Montemanni 1

1 Dalle Molle Institute for Artificial Intelligence (IDSIA – USI/SUPSI), Manno, Switzerland

roberto@idsia.ch

Abstract. Uncertainty is part of our everyday life. When optimizing a problem, it should therefore be taken into account. In the last 15 years great attention has been devoted to robust optimization, where the main effort has been in plugging uncertainty into the mathematical models describing the problems to optimize. A major challenge has been to treat uncertainty without (dramatically) increasing the computational complexity of the solving methods. However, most of the studies have been on the side of mathematical programming and exact methods, with only a very limited attention to heuristic approaches. In this talk we discuss how the results achieved for robustness in mathematical programming can be embedded directly into metaheuristic methods, obtaining the so-called matheuristic approaches. An important characteristic of these approaches is to be able to take uncertainty into account without a drastic increase in their computational requirements. The latter property is of great interest for practitioners since real-life problems cannot be normally treated by exact, or demanding methods in general, due to their large size. Extensions involving solutions pools and parallelization, topics again of great interest for practitioners, are finally discussed. The robust matheuristic framework will be illustrated with examples and results on vehicle routing problems.


APPLYING AN IMMUNE ANT COLONY SYSTEM ALGORITHM TO SOLVE AN INTEGRATED FLEXIBLE BAY FACILITY LAYOUT PROBLEM WITH INPUT/OUTPUT POINTS DESIGN

Yun-Zhu Lin 1 and Yu-Chin Lin 1

1 Department of Civil Engineering, Chung Yuan Christian University, Taoyuan City, Taiwan

{mschang, yuchi}@cycu.edu.tw

Abstract. Traditionally, most studies use a two-stage approach to solve a block layout problem with input/output points design. To make the planning results more practical and measure distance between facilities more precisely, this study integrates a flexible bay facility layout problem and input/output points design using a contour distance metric. In this study, an ant colony system (ACS), clonal selection algorithm (CSA) and shortest path algorithm are combined, and an immune ant colony system (IACS) algorithm is proposed to solve an unequal-area facility layout problem with input/output points design. Operations of CSA are embedded in the ACS to improve the solution quality of initial ant solutions and increase the differences between the ant solutions, and the search capability of IACS is thus enhanced. Nine international benchmark problems are used to test the algorithm efficiency of IACS. When compared to previous research, IACS can deliver new and better solutions.


FLEET SIZE AND MIX VEHICLE ROUTING PROBLEM WITH BACKHAULS: A MULTI-START APPROACH

Javier Belloso 1, Angel A. Juan 2, Javier Faulin 1 and Adrián Serrano 1

1 Public University of Navarra, 31006 Pamplona, Spain

{javier.belloso, javier.faulin, adrian.serrano}@unavarra.es

2 Open University of Catalonia, 08018 Barcelona, Spain

ajuanp@uoc.edu

Abstract. We consider the Fleet Mixed Vehicle Routing Problem with Backhauls (FSMVRPB), where delivery and pick-up customers are going to be served from a central depot and the fleet of vehicles is unlimited and heterogeneous. The proposed algorithm utilizes a multi-start approach to prevent the potential solution of being attracted by local minima. This method combines three randomized criteria, the first one for selecting a vehicle type and the second one for sorting the savings list as a result of applying an efficient method. Once the vehicle type is selected, the algorithm solves the homogeneous VRPB and promotes a number of routes to the heterogeneous solution using the third randomized criteria. For this selection, it is considered the range between 0 and the total number of routes solving the homogeneous problem. Benchmark instances for the FSMVRPB have been selected in order to assess the efficiency of our approach, and results show that our approach is able to provide promising solutions by improving some of the best solutions reported in the literature.


HYBRID-DEMAND QUEUEING FOR COMMUTER PARKING

Charles D. Pack 1

1 Monmouth University, 170 Hudson Ave., Red Bank, NJ USA

cpack@monmouth.edu

Abstract. Registered (finite-source) commuters and (infinite-source) visitors share a commuter parking lot. We develop a novel, hybrid-demand equilibrium queueing model, with demand dynamics, for use in traffic-engineering these lots. Our transient version can help analyze the daily lot startup process. Finally, we propose criteria and take initial steps to evaluate our models. While the problem is not new, there are essentially no other published analytical results or detailed studies regarding commuter lot engineering.


ESG PORTFOLIO OPTIMIZATION: INTEGRATING COMBINATORIAL GOAL PROGRAMMING AND CORPORATE RESPONSIBILITY RATINGS

Gordon H. Dash 1 and Nina Kajiji 1,2

1 University of Rhode Island, Kingston, USA

ghdash@uri.edu

2 The NKD-Group, Inc. USA

nina@nkd-group.com

Abstract. Investment approaches that embrace environmental, sustainability and governance factors (ESG factors) have shown evidence of providing investors with potential long-term investment performance advantages. With global approaches to investing becoming more predominant among various investment professionals, ESG factors provide portfolio managers with a wider view of a company's risk-return profile across systemic relationships. This paper introduces the combinatorial nonlinear multiple objective optimization model (MINLGP) of Dash and Kajiji (2014) to optimize a goal-oriented ESG portfolio based on the Thomson Reuters Corporate Responsibility Ratings. Temporal instability of correlation profiles is also treated by the proposed model. Following similar approaches in the literature, the MINLGP optimized portfolio is rebalanced periodically to incorporate new correlation information. During the interim periods between rebalancing periods, an active real-time futures-based hedging strategy is invoked to stabilize or enhance the desired risk-return outcomes. A temporal simulation of a MINLGP portfolio is undertaken by invoking the investment profile of an affiliate of the global non-profit service organization – World Association of Girl Guides and Girl Scouts (WAGGGS).


© ORLab Analytics Inc.