Home  << 1 >> 
Freire, A. S., Moreno, E., & Yushimito, W. F. (2016). A branchandbound algorithm for the maximum capture problem with random utilities. Eur. J. Oper. Res., 252(1), 204–212.
Abstract: The MAXIMUM CAPTURE PROBLEM WITH RANDOM UTILITIES seeks to locate new facilities in a competitive market such that the captured demand of users is maximized, assuming that each individual chooses among all available facilities according to the wellknow a random utility model namely the multinomial logit. The problem is complex mostly due to its integer nonlinear objective function. Currently, the most efficient approaches deal with this complexity by either using a nonlinear programing solver or reformulating the problem into a MixedInteger Linear Programing (MILP) model. In this paper, we show how the best MILP reformulation available in the literature can be strengthened by using tighter coefficients in some inequalities. We also introduce a new branchandbound algorithm based on a greedy approach for solving a relaxation of the original problem. Extensive computational experiments are presented, bench marking the proposed approach with other linear and nonlinear relaxations of the problem. The computational experiments show that our proposed algorithm is competitive with all other methods as there is no method which outperforms the others in all instances. We also show a largescale real instance of the problem, which comes from an application in parkandride facility location, where our proposed branchandbound algorithm was the most effective method for solving this type of problem. (C) 2015 Elsevier B.V. All rights reserved.

HolguinVeras, J., Yushimito, W. F., ArosVera, F., & Reilly, J. (2012). User rationality and optimal parkandride location under potential demand maximization. Transp. Res. Pt. BMethodol., 46(8), 949–970.
Abstract: The paper develops analytical formulations to gain insight into the optimal location, i.e., the one that maximizes the potential market, and to estimate the potential catchment area of Park and Ride (P&R) facilities. The formulations are based on the assumption that a traveler would use a P&R facility if and only if the corresponding generalized cost is lower than the drive only alternative. The paper considers two different scenarios: a linear city (or a travel corridor), and a twodimensional city with Euclidean travel. Analytical derivations were obtained for both cases using, as starting point, the necessary condition for P&R use. In the case of the linear city, the paper identifies two breakeven distances (BEDs) of great import to the estimation of the potential P&R market: the (trip) origin BED, i.e., the distance below which a traveler could drive upstream to use the P&R facility to access its downstream destination, and still be better off; and the (trip) destination BED, i.e., the travel distance using transit below which it does not make sense to use P&R. The paper proves that the optimal location of P&R sites is shifted upstream of what seems to be an intuitive solution, i.e., the edge of the congested region. by a distance that depends on the relative values of the origin and destination BEDs. In the twodimensional city case, the analytical derivations prove that, for a given trip from i to j, the set of feasible locations follows an ellipselike figure with the trip origin as a focus. These shapesreferred to as limiting functionsdepend on variables such as trip distance, transit level of service (LOS), and the like. The analyses indicate that the area enclosed by the limiting functions increases with the transit LOS and trip distance, and so do the corresponding catchment areas. This is because the catchment area is determined by the marginal trip origins, i.e., those for which the P&R facility is just inside the limiting function. In its final section, the paper develops a parabolic approximation to the catchment area for a given P&R site. The approximating parabola is defined by three critical points: the origin BED, and two points that identify the marginal trip origins at the chord of parabola evaluated at the P&R. The numerical experiments indicate that the parabolic approximation provides a fairly good estimate of the catchment area that is easy to produce, conceptually valid, and overcomes the limitations of alternative approaches and rules of thumb used by practitioners and researchers. (c) 2012 Elsevier Ltd. All rights reserved.
Keywords: Park and Ride; Optimal location; Transit

Hughes, S., Moreno, S., Yushimito, W. F., & HuertaCanepa, G. (2019). Evaluation of machine learning methodologies to predict stop delivery times from GPS data. Transp. Res. Pt. CEmerg. Technol., 109, 289–304.
Abstract: In last mile distribution, logistics companies typically arrange and plan their routes based on broad estimates of stop delivery times (i.e., the time spent at each stop to deliver goods to final receivers). If these estimates are not accurate, the level of service is degraded, as the promised time window may not be satisfied. The purpose of this work is to assess the feasibility of machine learning techniques to predict stop delivery times. This is done by testing a wide range of machine learning techniques (including different types of ensembles) to (1) predict the stop delivery time and (2) to determine whether the total stop delivery time will exceed a predefined time threshold (classification approach). For the assessment, all models are trained using information generated from GPS data collected in Medellin, Colombia and compared to hazard duration models. The results are threefold. First, the assessment shows that regressionbased machine learning approaches are not better than conventional hazard duration models concerning absolute errors of the prediction of the stop delivery times. Second, when the problem is addressed by a classification scheme in which the prediction is aimed to guide whether a stop time will exceed a predefined time, a basic Knearestneighbor model outperforms hazard duration models and other machine learning techniques both in accuracy and F1 score (harmonic mean between precision and recall). Third, the prediction of the exact duration can be improved by combining the classifiers and prediction models or hazard duration models in a two level scheme (first classification then prediction). However, the improvement depends largely on the correct classification (first level).
Keywords: Machine learning; Stop delivery time; Classification; Regression; Hazard duration; GPS

Silva, K., Lima, R. D., Alves, R., Yushimito, W. F., & HolguinVeras, J. (2020). Freight and Service Parking Needs in Historical Centers: A Case Study in Sao Joao Del Rei, Brazil. Transp. Res. Record, 2674(11), 352–366.
Abstract: The objective of this study is to analyze the demand for loading and unloading parking spaces in the center of Sao Joao Del Rei, a historical city in the State of Minas Gerais, Brazil, through freight trip generation models. To generate the models, the number of employees is used as an independent variable. Results show that the historical center receives an average of 710 freight trips per day, which would require at least 43 spaces for loading and unloading. As the center has only eight such spaces available, representing 18% of total demand, this study proposes new locations and suggests transportation demand management measures that could be used in conjunction with the allocation of new parking spots.

Yushimito, W. F., Alves, P. N., Canessa, E., & de Mateo, F. (2018). Relating Efficiency With Service Compliance Indices In Public Transportation Using SlackBased Measure Data Envelopment Analysis And Shadow Prices. Promet, 30(6), 661–670.
Abstract: In many countries, bus operators are private companies whose service has been leased by government agencies. These agencies develop service compliance indices or measures to keep track of factors such as passenger satisfaction, frequency, and regularity but do not necessarily include the objectives of the operators in the assessment. In this paper, we used slackbased measure data envelopment analysis (SBM) to investigate whether it is possible for a bus operator to be efficient (from a private perspective) and match required standards of frequency and regularity. In doing so, data collected from two major bus operators in Santiago, Chile has been used comprising 99 services. The results show that when private objectives, namely revenues, are included in the analysis, bus operators do not necessarily seek to improve the regularity of their service. Moreover, it was found that some bus services are on the efficient frontier while keeping low performance measure standards. Using the shadow prices of the models, it was also found that improving the performance measures will be hard for many bus services unless there is a significant change in factors that are not under control of the operators (i.e., number of stops, length of the route, etc.). This shows the difficulty of correctly aligning the private objectives of operators with agencies' objectives.

Yushimito, W. F., Ban, X. G., & HolguinVeras, J. (2014). A TwoStage Optimization Model for Staggered Work Hours. J. Intell. Transport. Syst., 18(4), 410–425.
Abstract: Traditional or standard work schedules refer to the requirement that workers must be at work the same days and during the same hours each day. This requirement constrains workrelated trip arrivals, and generates morning and afternoon peak hours due to the concentration of work days and/or work hours. Alternative work schedules seek to reschedule work activities away from this traditional requirement. The aim is to flatten the peak hours by spreading the demand (i.e., assigning it to the shoulders of the peak hour), lowering the peak demand. This not only would reduce societal costs but also can help to minimize the physical requirements. In this article, a twostage optimization model is presented to quantify the effects of staggered work hours under incentive policies. In the first stage, a variation of the generalized quadratic assignment problem is used to represent the firm's assignment of workers to different work starting times. This is the input of a nonlinear complementarity problem that captures the behavior of the users of the transportation network who are seeking to overcome the constraints imposed by working schedules (arrival times). Two examples are provided to show how the model can be used to (a) quantify the effects and response of the firm to external incentives and (b) evaluate what type of arrangements in starting times are to be made in order to achieve a social optimum.

Yushimito, W. F., Ban, X. G., & HolguinVeras, J. (2015). Correcting the Market Failure in Work Trips with Work Rescheduling: An Analysis Using Bilevel Models for the Firmworkers Interplay. Netw Spat. Econ., 15(3), 883–915.
Abstract: Compulsory trips (e.g., work trips) contribute with the major part of the congestion in the morning peak. It also prevents the society to reach a social optimum (the solution that maximizes welfare) because the presence of the private utility of one the agents (the firm), acting as a dominant agent, does not account for the additional costs imposed in their workers (congestion) as well as the costs imposed to the rest of the society (i.e., congestion, pollution). In this paper, a study of a strategy to influence the demand generator by relaxing the arrival constraints is presented. Bilevel programming models are used to investigate the equilibrium reached from the firmworkers interplay which helps to explain how the market failure arises. The evaluation includes the use of incentives to induce the shift to less congested periods and the case of the social system optimum in which a planner objective is incorporated as a third agent usually seeking to improve social welfare (improve productivity of the firm while at the same time reduce the total system travel time). The later is used to show that it is possible to provide a more efficient solution which better off society. A numerical example is used to (1) show the nature of the market failure, (2) evaluate the social system optimum, and (3) show how a congestion tax or an optimal incentive can help to correct the market failure. The results also corroborate that these mechanisms are more likely to be more efficient when firms face little production effects on time and workers do not high opportunity costs for starting at off peak periods.

Yushimito, W. F., HolguinVeras, J., & Gellona, T. (2018). Firm's efficiency and the feasibility of incentives for flextime adoption: a preliminary analysis of Chilean employer's response. Transp. Lett., 10(4), 202–214.
Abstract: Flextime is a Travel Demand Management policy that allows workers to arrive/leave within a certain threshold of time but staying on duty during a period of the day (core hours). Such flexibility in arrival time can have an impact on relieving traffic congestion as it can shift demand to nonpeak periods. However, in several countries, companies are usually reluctant to provide flexibility arguing possible loss in their productivity and efficiency. This paper discusses a preliminary study on perceptions of challenges, and the potential of flextime in congested urban areas. In doing so, we present the information gathered through indepthinterviews and subsequent questionnaires applied to managers of cars dealers in Santiago, Chile, obtaining both qualitative and quantitative information. The quantitative information obtained was used to develop (1) a data envelopment analysis to investigate whether firms' productivity is affected by flextime or not, and (2) a binary logit model to evaluate financial and nonfinancial incentives that can allow firms to adopt flextime or change work arrival times. The results obtained from the sample show that, for this particular industry, flextime might not have a significant impact on firms' productivity. Additionally, we found that financial incentives and competitors' work starting times have a significant impact on accepting flextime or relaxing arrival constraints. We also found that if arrangements with the whole industry to change work starting times are made, the effect of this change can be as effective as the provision of financial incentives.
