Papers by Denis Borenstein
Pesquisa Operacional, Dec 31, 2022
Pesquisa Operacional
This work considers the strategic Supply Chain Network Design (SCND) problem, which is to define ... more This work considers the strategic Supply Chain Network Design (SCND) problem, which is to define the number and location of facilities, and the flow of products among them to fulfilling a long-term deterministic demand. A two-phase heuristic approach was specially developed to solve large scale problems in reasonable time, extending a previous algorithm introduced in Farias et al. (2017). In the construction phase, a multi-start approach was developed to generate diversified initial solutions from each new iteration of a layered-based rounding heuristic. In the second phase, a local search heuristic improves the solution provided by the rounding method. The solution method is evaluated using randomly generated instances, and a evaluated strategic of marketing in a real case study applied to a company to redesigning the supply chain to two lines of products.The obtained results evidence the effectiveness and flexibility of the developed approach for handling very large instances.
Journal of Retailing
Abstract Forecasting is one of the fundamental inputs to support planning decisions in retail cha... more Abstract Forecasting is one of the fundamental inputs to support planning decisions in retail chains. Frequently, forecasting systems in retail are based on Gaussian models, which may be highly unrealistic when considering daily retail data. In addition, the majority of these systems rely on point forecasts, limiting their practical use in retailing decisions, which often requires the full predictive density for decision making. The main contribution of this paper is the modeling of daily distribution centers (DCs) level aggregate demand forecasting using a recently proposed framework for non-Gaussian time series called score-driven models or Generalized Autoregressive Score (GAS) models. An experimental study was carried out using real data from a large retail chain in Brazil. A log-normal GAS model is compared to usual benchmarks, namely neural networks, linear regression, and exponential smoothing. The results show the GAS model is a competitive alternative to retail demand forecasting in daily frequency, with the advantage of producing a closed form predictive density by construction.
Anais do I Simpósio Brasileiro de Sistemas de Informação (SBSI 2004), 2004
O artigo apresenta um sistema de apoio à decisão aplicado ao planejamento operacional da coleta s... more O artigo apresenta um sistema de apoio à decisão aplicado ao planejamento operacional da coleta seletiva de resíduos sólidos (SCOLDSS), o qual tem por funcionalidade principal à geração de alternativas ao processo decisório no que se refere á: (a) alocação de veiculos para a roleta seletiva, bem como o roteiro a ser percorrido pelos mesmos e, (b} a determinação da quantidade diária de resíduos sólidos a ser enviado a cada unidade de triagem. Para o desenvolvimento do mesmo foi utilizada a combinação de técnicas advindas da Pesquisa Operacional, que são a simulação computacional de eventos discretos e algoritmos para o problema da alocação e roteamento de veículos. O sistema foi desenvolvido utilizando o ambiente Borland Delphi e, para a simulação foi utilizado o simulador Arena 3.5. Para a validação do SCOLDSS estão sendo utilizados dados da coleta seletiva de um município do Rio Grande do Sul.
Finance Research Letters, 2018
In this paper, we compare risk measures regarding performance of optimal portfolio strategies. We... more In this paper, we compare risk measures regarding performance of optimal portfolio strategies. We consider eleven risk measures from different classes. In particular, we propose a formulation that generates from any loss measure, a deviation based on the dispersion of results worse than it, which leads to very interesting risk measures. We consider 198,000 portfolios composed by stocks of the U.S. equity market, considering different scenarios in a simulation framework. Results indicate there is no clearly dominant risk measure. Despite this lack of dominance, including deviation terms consistently exhibits advantages regarding performance.
Transportation Research Part B: Methodological, 2018
The multiple-depot vehicle type rescheduling problem (MDVTRSP) is a dynamic extension of the clas... more The multiple-depot vehicle type rescheduling problem (MDVTRSP) is a dynamic extension of the classic multiple-depot vehicle scheduling problem (MDVSP), where a heterogeneous fleet is considered. The MDVTRSP consists of finding a new schedule given that a severe disruption occurred in previously scheduled trips very quickly, simultaneously minimizing the transportation costs and the deviations from the original plan. Although several mathematical formulations and solution methods have been developed for the robust MDVTRSP, the real time MDVTRSP is still unexplored. In this paper, we introduce a formulation of the problem and develop a heuristic solution method, employing time-space network, truncated column generation, and preprocessing procedures. The solution method has been implemented in several algorithm variants, combining different developed preprocessing methods. Computational experiments on randomly generated instances were performed to evaluate the performance of the developed algorithms. The best solutions concerning efficiency and efficacy were obtained by the variants considering state space reductions to accelerate the convergence process of the column generation. Solutions were obtained very quickly (in less than 150 seconds for large instances, considering up to 2500 trips, eight depots, and one breakdown. The developed heuristics also presented a good behavior for several simultaneous disruptions, solving the problem with a little increase (less than 8.5%, on average) in the required CPU time. A case study using data from a real-life small instance in Brazil also demonstrated the efficiency and efficacy of the approach when compared with manual planning strategies.
SSRN Electronic Journal, 2017
Using sectorial indices of the Brazilian market, we compare the portfolio optimization approach k... more Using sectorial indices of the Brazilian market, we compare the portfolio optimization approach known as risk parity with minimum variance and equally weighted approaches. We apply various estimators for the covariance matrix to each portfolio strategy, since portfolio variance is considered as risk measure. Empirical results demonstrate that the risk parity approach provides more diversified portfolios and stable weights in the out-of-sample than the other two approaches, thereby avoiding the dangers of excessive concentration and reducing transaction costs. Furthermore, the results demonstrate that different estimators of the covariance matrix had little influence on the results obtained through the risk parity approach.
Energy, 2016
Este documento contiene información de prueba. Contáctese con el administrador del Centro para el... more Este documento contiene información de prueba. Contáctese con el administrador del Centro para el acceso al documento originar del registro.
Este trabalho de pesquisa tem como objetivo desenvolver uma modelagem de Sistemas Flexíveis de Ma... more Este trabalho de pesquisa tem como objetivo desenvolver uma modelagem de Sistemas Flexíveis de Manufatura (FMS), fornecendo a administradores e engenheiros uma ferramenta objetiva de análise a ser utilizada durante o projeto e operação destes sistemas. Essa ferramenta é um modelo genérico de simulação computacional capaz de replicar, em laboratório, os fenômenos que ocorrem no interior de sistemas discretos de produção'em lotes. Desta forma, permite a esses profissionais que conheçam e questionem os mesmos, reduzindo os riscos e incertezas inerentes ao processo de implementação prática. O trabalho apresenta uma discussão do conceito de FMS, explicitando a necessidade de uma ferramenta de análise, bem como uma descrição completa do modelo abstrato desenvolvido e de sua representação computacional, denominada de SIMFLEX.The aim of this reseach is to model a Flexible Manufacturing Systems(FMS), offering to managers and engineers an objective tool of analysis to be used during the design and operation of these systems. The tool is a computational simulation model that emulates, in laboratory, the events presented in batch production systems. It allows a better knowledge and inquiring of these systems by these professionals, decreasing'risks and uncertanties inherent in theirs implementation process. The work presents a discussion of the FMS concept, emphasizing the need of a simulation tool such as this. A complete description of the abstract model developed and its computational representation, called SIMFLEX, is also presented
Engineering Applications of Artificial Intelligence, 2016
Este documento contiene información de prueba. Contáctese con el administrador del Centro para el... more Este documento contiene información de prueba. Contáctese con el administrador del Centro para el acceso al documento originar del registro.
Annals of Information Systems, 2007
This paper presents the conception, modeling, and implementation of a decision support system app... more This paper presents the conception, modeling, and implementation of a decision support system applied to the operational planning of solid waste collection systems, called SCOLDSS. The main functionality of the system is the generation of alternatives to the decision processes concerning: (a) the allocation of separate collection vehicles, as well as the determination of their routes, and (b) the determination
Periódico Eletrônico Fórum Ambiental da Alta Paulista, 2012
This paper presents the formulation and the computational solution of a procedure for ranking alt... more This paper presents the formulation and the computational solution of a procedure for ranking alternatives in Fuzzy-Electre environments. It attempts to exploit the well-known advantages of the Fuzzy-TOPSIS method in order to define a new ranking procedure by combining the pure concordance and discordance index with the ideal positive and negative solutions. The final ranking of the alternatives is given by a modified closeness coefficient. A case study related to a purchasing situation in a hospital company is presented and discussed in details. Results obtained show that the approach developed in this work has potential as an alternative ranking procedure to Fuzzy-Electre method.
Kliewer et al. (2006b, 2011) proposed a time window implementation to the multi-depot vehicle (an... more Kliewer et al. (2006b, 2011) proposed a time window implementation to the multi-depot vehicle (and crew) scheduling problem using a time-space network (TSN). Based on this approach, we developed a new methodology to implement time windows to the vehicle-type scheduling problem (VTSP), which solves the vehicle scheduling problem considering heterogeneous fleet. Our method presents as main advantages the simplicity in the implementation, a smaller sized network, and facility to introduce new constraints closer to reality. In order to verify the effectiveness of our approach, experiments were carried out using real instances from a Brazilian city and large random instances. Analyzing the obtained results, it is possible to affirm that the developed method is able to present relevant savings in the daily operations of the public transportation service, reducing the required number of scheduled vehicles to satisfy the historic demand.
Lecture Notes in Economics and Mathematical Systems
When a bus on a scheduled trip breaks down, one or more buses need to be rescheduled to serve the... more When a bus on a scheduled trip breaks down, one or more buses need to be rescheduled to serve the customers on that trip with minimum operating and delay costs. The problem of reassigning buses in real-time to this cut trip, as well as to other scheduled trips with given starting and ending times, is referred to as the bus rescheduling problem (BRP). This paper considers modeling, algorithmic, and computational aspects of the single-depot BRP. The paper develops the sequential and parallel auction algorithm to solve the BRP. Computational results show that our approach solves the problem quickly.
Waste Management, 2007
This study presents the conception, modeling, and implementation of a decision support system app... more This study presents the conception, modeling, and implementation of a decision support system applied to the operational planning of solid waste collection systems, called SCOLDSS. The main functionality of the system is the generation of alternatives to the decision processes concerning: (a) the allocation of separate collection vehicles, as well as the determination of their routes and (b) the determination of the daily amount of solid waste to be sent to each sorting unit, in order to avoid waste of labor force and to reduce the amount of waste sent to the landfills. To develop the computer system, a combination of quantitative techniques was used, such as: simulation of discrete events and algorithms/heuristics for vehicle allocation and routing. The system was developed using the Borland Delphi environment and the commercial software Arena to carry out the simulations. We also present a computational study with real-life data from the solid waste collection in Porto Alegre, Brazil, in which we show that the results provided by the computational system outperform the operation planning currently adopted.
Omega, 2008
This paper considers a truck scheduling problem in the context of solid waste collection in the C... more This paper considers a truck scheduling problem in the context of solid waste collection in the City of Porto Alegre, Brazil. The problem consists of designing "good" daily truck schedules over a set of previously defined collection trips, on which the trucks collect solid waste in fixed routes and empty loads in one of several operational recycling facilities in the system. These facilities are managed by cooperatives whose members are poor and not part of the mainstream economy. The main objective is to minimize the total operating and fixed truck costs. We show that the problem can be modeled as a special case of the single-depot vehicle scheduling problem, which is polynomially solvable. However, due to the social benefits of the solid waste program, it is desirable to obtain balanced assignments of collection trips unloading their cargo at the recycling facilities. We prove that the truck scheduling problem considering balanced unloading is NP-hard. A heuristic approach, incorporating an auction algorithm and a dynamic penalty method, is designed to acquire a good solution. Finally, computational experiments are conducted on real data. The results show that the heuristic approach simultaneously reduces total costs and balances the number of trips assigned to each recycling facility.
Networks, 2007
When a vehicle on a scheduled trip breaks down, one or more vehicles need to be rescheduled to se... more When a vehicle on a scheduled trip breaks down, one or more vehicles need to be rescheduled to serve the customers on that trip with minimum operating and delay costs. The problem of reassigning vehicles in real‐time to this cut trip as well as to other scheduled trips with given starting and ending times, is referred to as the vehicle rescheduling problem (VRSP). This paper considers modeling, algorithmic, and computational aspects of the single‐depot VRSP. The paper formulates a model for this problem and develops several fast algorithms to solve it, including parallel synchronous auction algorithms. The concept of the common feasible network (CFN) is introduced to find a good set of initial “prices” for speeding up the auction algorithm. Computational experiments on randomly generated problems are described. Computational results show that, for small problems, all of the developed algorithms demonstrate very good computational performances. For large problems, parallel CFN‐based ...
Expert Systems with Applications, 2012
The selection supplier problem has received a lot of attention from academics in recent years. Se... more The selection supplier problem has received a lot of attention from academics in recent years. Several models were developed in the literature, combining consolidated operations research and artificial intelligence methods and techniques. However, the tools presented in the literature neglected learning and adaptation, since this decision making process is approached as a static one rather than a highly dynamic process. Delays, lack of capacity, quality related issues are common examples of dynamic aspects that have a direct impact on long-term relationships with suppliers. This paper presents a novel method based on the integration of influence diagram and fuzzy logic to rank and evaluate suppliers. The model was developed to support managers in exploring the strengths and weaknesses of each alternative, to assist the setting of priorities between conflicting criteria, to study the sensitivity of the behavior of alternatives to changes in underlying decision situations, and finally to identify a preferred course of action. To be effective, the computational implementation of the method was embedded into an information system that includes several functionalities such as supply chain simulation and supplier's databases. A case study in the biodiesel supply chain illustrates the effectiveness of the developed method.
Uploads
Papers by Denis Borenstein