Operations Research and Analytics in Latin America

Vehicle congestion in Bogota is high, generating high operating costs and long waiting times for school buses for children. Currently, in Bogota, and in most of Colombian cities, each school owns its own transportation fleet or outsources buses for the exclusive transportation service for its own students. Student pick-up in the morning and drop-off in the afternoons can be each one modeled as the well-known bus routing problem, which is a special case of the vehicle routing problem (VRP) with a single depot (the school). This approach of the problem can generate inefficiencies when analyzing schools located nearby within the same area. Therefore, this paper compares the results obtained by applying traditional routing models, where each school picks up its students, with a collaborative transportation scheme, in which students from different schools are assigned to a shared bus fleet. This last approach allows schools in the same area picking up students regardless of the school they belong to. Mathematical models are proposed and solved using a commercial solver. As a result, due to the number of pick-up points for each configuration, the solver was not able to reach the global optimum for any of both “traditional” and “collaborative” routes. Despite this, the collaborative transportation model showed advantages in terms of the total distance traveled by the bus fleet, which decreases by 4.37%. The paper opens opportunities for further exploring the collaboration bus routing problem.

This document studies one of the many applications and challenges that call centers face for staff scheduling. The objective of this document is to design and implement a solution technique that seeks to minimize costs due to missing teleoperators during lunch and breaks time for different operations of teleoperators with pre-defined shifts, reducing execution times and fulfilling the restrictions that each operation has. As part of the limitations identified for the development of this simulation, it was established that teleoperators have a defined shift with start and end time, according to the labor contract of each teleoperator, and the scenarios were simulated in 24-h operation as maximum, to guarantee one tool which responds to possible daily variations that may arise in different scenarios. In this context, genetic algorithm was proposed to provide a solution regarding to all requirements, achieving results near to the optimum. Besides, we defined the execution of four scenarios composed of operations with 200, 500, 2000 and 5000 teleoperators, each with different constraints and variables, to evaluate how the proposed algorithm responds. As result, solutions were generated with better performance in comparison with company’s actual scheduling model for lunch and break times, achieving on average a 42% decrease in total cost and 69% reduction in tool execution time. Also, we got better results for computational time in scenarios with 200 and 500 teleoperators, showing 76% less compared to current company’s tool, and for instances with more teleoperators (2000 and 5000), better results are presented with 60% reduction of costs delivered by the company’s tool.

The increased popularity of specialized dairy farms, mainly in the Colombian Andean region, represents an investment opportunity and a contribution to national agricultural growth. In this way, the effort to strengthen the livestock legacy, especially by young entrepreneurs, focuses on technified facilities that bet on improving production conditions for farmers, but at the same time, that contribute to the development of the local dairy industry. On the other hand, the dealers of dairy products, together with government entities, have been improving conditions for their producers through financial facilities and incentives for good livestock practices. It seeks to encourage interest in investment in Colombian agriculture, especially in dairy farms. This work aimed to evaluate the financial feasibility of the installation and operation of a specialized dairy farm in Colombia. Based on projected cash flows, financial evaluation criteria were calculated for two credit line scenarios. Additionally, through the use of the RiskSimulator simulation software, risk analysis, and a sensitivity analysis were carried out to evaluate the cash flows from the investor's perspective, as well as to identify the critical variables for the success of the project. The results indicated the feasibility of implementing the dairy unit from the investor's perspective, using the government financing option (Finagro). It was also identified that the costs related to cattle feeding, the sale price of milk, and the amount of milk produced are the variables to which the project indicators are most sensitive. Therefore, they are generating the greatest source of volatility in the project.

A positioning analysis was performed to 8 different brands of automotive lubricants in Colombia. For the analysis, information was collected on the perception that the interviewees (managers of automotive lubricant sales points) had about 13 characteristics of the brands along with their perception about what they considered to be an ideal brand. From the information, graphic tools known as perceptual maps were produced, allowing us to identify the best perceived attributes of the brands, their similarities and differences with the ideal brand and the brands that compete with each other. Features such as warranty, product availability, discounts, and a focus on heavy-duty products are associated with an ideal brand, while attributes related to product quality and performance are associated with traditional brands in the market.

Attention to scheduling problems with learning effect has increased, since the factors that influence productivity of manual tasks are being considered recently. The flowshop system is one of the most frequent configurations of hand-intensive production systems; it belongs to the class of NP-hard combinatorial optimization problems. Thus, this article develops an algorithm to resolve the flowshop scheduling problem with learning effect with makespan minimization. Four models for calculating the learning effect referred to in the literature are considered (according to the position and the sum of the processing times). This paper proposes the Nawaz-Enscore-Ham Algorithm (NEH) with two local search operators. This algorithm is tested through computer experiments.

This paper describes the stages for the formulation of a transdisciplinary productivity model based on the integration of previously sele1cted exact, social, and human variables in doctoral research. The set of social factors are classified according to the identity of people and their interaction in a group environment. Exact variables are related to the processes, their resources, and the government. Soft sciences variables constitute the differential of the construct with respect to the traditional models of productivity measurement, structured based on exact variables. Stages form the developed methodology. Results include obtaining the relevant variables from the correlation of the aggregate values of the applied validation and the formulation of the model protocol. The mathematical modeling is under development.

Over the last years, there have been significant technological advances in electric vertical take-off and landing (eVTOL) vehicles. Currently, multiple companies are working to establish new air transportation services in urban areas, such as air taxis. With the growing number of projects and interest in these new air services, urban airspace can become congested within the next decade. A key element of these emerging aviation markets that uses eVTOL vehicles is their take-off and landing areas, known as vertiports or skyports. The high volume of expected eVTOL aircraft can cause vertiports to become the bottleneck of urban air operations. Therefore, the aircraft should be efficiently scheduled using appropriate decision tools. This study proposes two mathematical formulations for scheduling take-off and landing aircraft at vertiports. The models consider urban air mobility operational characteristics such as separation times and blocking constraints to avoid collisions while the aircraft use the vertiport. As objective function, we minimize the completion of the last aircraft operation at vertiports. Results show that the optimization models assign and sequence aircraft at vertiport components to both departing and landing aircraft considering separations in the vertiport gates, touchdown and liftoff (TLOF) pads, and in common ground taxi routes. This paper hence opens avenues for further research.

This paper presents a mathematical model of a potato supply and distribution chain in Latin America. We formulate a model for the distribution process from a set of potato producers to a set of nanostores (small retailers, common in Latin American countries) in both Bogotá, Colombia, and Cochabamba, Bolivia. In this model, we located a set of warehouses to consolidate and facilitate the distribution. The objective function was to minimize the sum of all costs subject to capacity, demand satisfaction, and coverage constraints. This model was implemented on the GAMS software; the results were translated into valuable information used to give shape to the distribution chain and to perform an analysis to further understand how the change in some of the input data could affect the optimal solution. The results show that even though the model implemented in both countries is the same, the results obtained are widely different because of each country's characteristics. This research opens the doors for future research that might consider traffic in the cities, criteria for the availability of producers, and even a vehicle routing solution for the delivery trucks. This model was developed jointly by students of Universidad de La Sabana and of Universidad Privada Boliviana as a part of a COIL (Collaborative Online International Learning) project.

In this paper, we consider the problem of prioritization of road reconstruction in a disaster-affected rural network. We propose a solution based on a labeled network that aims to maximize the accessibility to victim locations. The proposed solution approach uses a labeled network in which the edges have an assigned priority. The labeled network refers to the road network, where each road edge has a value representing the road prioritization for recovery. We propose two criteria, road travel time and connectivity, to determine which roads should first be recovered to improve road access to victim locations. We present numerical studies using artificial instances with different disruption levels. The results indicate that the best prioritization criteria depend on the location of the disaster management center and disruption level. In addition, this paper provides insights on using flow network properties to design criteria for considering the structure of the damaged road network.

Humanitarian operations are characterized by the need to provide aid to affected populations in an efficient manner for minimizing the effects caused by humanitarian crises. Humanitarian operations include the distribution of aid and scarce resources such as water, medicine, and food, with the total cost of operation being one of the criteria most used by decision-makers and researchers. However, in the last two decades, the inclusion of social costs such as the deprivation cost and equity as part of the decision criteria in these problems has become a trend of great relevance. This research presents a heuristic method to establish a drinking water distribution plan in post-disaster scenarios including the estimation of the deprivation cost experienced by the people affected by humanitarian disasters. Twenty instances are solved based on available information from the department of Cundinamarca, Colombia. Results show superior performance on the drinking water distribution cost by the proposed method versus a commercial optimizer.

Emergency departments in hospitals are a crucial part of the healthcare systems around the world. They provide emergency medical services to patients who need urgent treatments. Due to the multiple sources of variability and uncertainty in their operations, they are considered complex systems. In addition, since the onset of COVID-19, hospitals have adapted their operations to decrease the contact between patients and medical staff to minimize the infection rate. In this work, we study the operations of an emergency department which made changes in its processes as a strategy to face the COVID-19 pandemic. The changes include the incorporation of a new area for respiratory patients, reallocating resources, and the modification of the patients’ journey in the emergency department. We propose a discrete event simulation approach to represent the operations before and after the changes in the emergency department services. The proposal was developed with collaboration of the emergency department managers and validated using the performance measures before the start of the COVID-19 pandemic. The main computational results show that the changes in the emergency department were effective to distribute the limited resources and limit the medical staff and non-respiratory patients’ exposure to suspected COVID-19 patients. However, in a post-COVID-19 scenario, such differentiation is no longer effective when the percentage of respiratory patients is less than 30%, since it increases the patients’ wait times, mainly for the non-urgent patients classified with less critical triage.

Hospital logistics represents an important role for the health sector, since it allows the optimization of resources in relation to the different key actors such as laboratories, distributors and hospitals, where challenges such as demand planning, inventories and maintenance, distribution in the production cycle, control of production and administration centers, implementation of a logistics area, among others, are presented. According to the above and after a search of information for the characterization of the health sector in Colombia, it was determined to carry out a logistic maturity model focused on hospitals, from the results obtained it was found that in general the sector is at level three (3) out of five in total. Thus, the document presents a consolidation of the previous phases, with a prioritization of strategies for the development of a roadmap to be used as a guide for the actors, thus improving their processes to reach a higher level of maturity.

Today's hospital systems are looking more than ever to make improvements in their internal processes to provide the best possible care for patients in their facilities and are turning to various management techniques such as Lean Healthcare and Analytics. This study describes the process of analyzing data about activity times and wastes for medicine administration, which is one of the most important activities of nursing care, and the construction of a predictive model. The methodology carried out in this research is a data collection of the medicines administration process and the patient to whom they are administered, observing whether there is waste or not. Subsequently, this information is integrated into a linear regression model for the estimation of working time, validating the corresponding assumptions. All this in order to develop better management in hospital wards. Regression models were constructed from a sample collected in the hospital wards with the epidemiological profile of the patients, and it is shown that the variables of the same come to generate an impact on the execution of the work. Finally, for future work, these models can be refined by including variables from the care environment and other activities, validated by professional nurses and linked to moments of nursing care.

This paper addresses the problem of vaccination and distribution that has brought COVID-19 in rural areas from Colombia. Locations far from main cities are the last considered on vaccines supply taking into account the distance and the access. There have been territories where vaccines took to arrive more than 40 days, making it hard for inhabitants from these regions to protect themselves on time from the pandemic. With this in mind, mass vaccination programs are needed in order to get to each region and guarantee the demand coverage. For this reason, it was decided to plan the location and distribution of vaccines to 200 potential sites in five different departments in Colombia, by means of a linear model developed in GAMS software, taking into account demand parameters, costs and transportation distances in order to obtain information to define vaccination points with their coverage areas and the amount of vaccines sent to each site, considering the main percentage of vaccination according to the population of each department in order to supply a large part of it, prioritizing people who do not have any dose and to those whose health condition requires it. This paper proposes a mathematical model based on demand satisfaction, distances, and capacity to select potential vaccination centers. In this document, first, the state of art is presented regarding the different approaches that have been developed to address this problematic. Second, an explanation of the problematic is provided choosing five regions in Colombia, especially rural areas where access to vaccination is limited. Third, the parameters used like distance, costs of transportation and operational costs for each vaccine, quantity of demand, vaccination capacity, and development of the model considering the objective function and constraints is explained. Fourth, a MILP optimization model approach is considered, and the results are analyzed. Finally, conclusions are provided as well as future research. With the results of the model, a possible solution is obtained suggesting that a total of 101 potential vaccine sites are required to satisfy the demand in the regions contemplated.

Global warming and climate change have far-reaching effects on our planet and have become the defining issues of our time. As one of the main polluters on the planet, the construction industry has a major role to play in terms of environmental impact reduction practices. This work intends to contribute to combat those serious impacts on the environment and make our cities more sustainable and oriented to achieve the Sustainable Development Goals. The present study assesses the feasibility of a housing project with sustainable parameters from a builder's perspective. Through forecasted cash flows, we compared a building already constructed using conventional systems with a building project of the same structure, but incorporating those sustainable parameters identified as the most important by the LEED standards and potential customers. Despite the relatively higher costs of sustainable construction, results indicated the viability of the project. Additionally, a factorial structure based on exploratory and confirmatory analyses is provided to evaluate two surveys used to determine those relevant sustainable factors which generate the greatest positive impact, and which are more valued by buyers. This statistical approach allowed the factors (latent variables) to be extracted in each of the two data samples, and it was found that the factorial structure reflects the underlying dimensions in accordance with the study results. Finally, our technical and financial studies demonstrate that green projects are feasible and there are opportunities to develop the green housing market.

Agriculture is experimenting economic and environmental challenges and the need to make decisions on improving performance and use of resources reductions as a priority. It is hence necessary to implement tools to aid decisions at strategic, tactical and operational levels. The current paper reviews existent literature with an emphasized on crop production systems and agricultural chain management, including planning, production, inventory management, and conditions of transport and distribution. The focus was planning, production techniques, operations and resources management. In contrast to narrative reviews, this paper follows a systematic protocol to avoid bias and allows replicability. Findings show that a high percentage of works emphases on hypothetical Operational Research (OR) models and optimization techniques that simplify the real problem with strong assumptions on isolated applied cases. These do not consider the integration of production, quality, food security, sustainability and risk conditions of a supply chain and many works only have been oriented to one echelon of the chain. Finally, the studies on agricultural production systems are based on some mathematical programming approaches, heuristics and artificial intelligence, but several are not applied or implemented in agricultural production. This paper also identifies key challenges and proposes opportunities for future research.

Venezuelans in an irregular migratory situation in Bogota must perform informal work that allows them to survive with their families. For this reason, this article presents a study of informal recyclers in Barrio La Concepción, through the method of problem solving and the implementation of surveys and observations in this neighborhood during January to May 2022. It was showed that informal recyclers are unaware of the migration regularization process to access the Temporary Protection Permit, as well as the labor formalization process to be registered as formal recyclers in the RURO. This lack of knowledge and the lack of state aid violate their fundamental rights and accentuate their unsatisfied basic needs.