Abstract
In sectorization problems, a large district is split into small ones, usually meeting certain criteria. In this study, at first, two single-objective integer programming models for sectorization are presented. Models contain sector centers and customers, which are known beforehand. Sectors are established by assigning a subset of customers to each center, regarding objective functions like equilibrium and compactness. Pulp and Pyomo libraries available in Python are utilised to solve related benchmarks. The problems are then solved using a genetic algorithm available in Pymoo, which is a library in Python that contains evolutionary algorithms. Furthermore, the multi-objective versions of the models are solved with NSGA-II and RNSGA-II from Pymoo. A comparison is made among solution approaches. Between solvers, Gurobi performs better, while in the case of setting proper parameters and operators the evolutionary algorithm in Pymoo is better in terms of solution time, particularly for larger benchmarks. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
Sectorization problems have significant challenges arising from the many objectives that must be optimised simultaneously. Several methods exist to deal with these many-objective optimisation problems, but each has its limitations. This paper analyses an application of Preference Inspired Co-Evolutionary Algorithms, with goal vectors (PICEA-g) to sectorization problems. The method is tested on instances of different size difficulty levels and various configurations for mutation rate and population number. The main purpose is to find the best configuration for PICEA-g to solve sectorization problems. Performance metrics are used to evaluate these configurations regarding the solutions’ spread, convergence, and diversity in the solution space. Several test trials showed that big and medium-sized instances perform better with low mutation rates and large population sizes. The opposite is valid for the small size instances. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
Sectorization is the process of grouping a set of previously defined basic units (points or small areas) into a fixed number of sectors. Sectorization is also known in the literature as districting or territory design, and is usually performed to optimize one or more criteria regarding the geographic characteristics of the territory and the planning purposes of sectors. The most common criteria are equilibrium, compactness and contiguity, which can be measured in many ways. Sectorization is similar to clustering but with a different motivation. Both aggregate smaller units into groups. But, while clustering strives for inner similarity of data, sectorization aims at outer homogeneity [1]. In clustering, groups should be very different from each other, and similar points are classified in the same cluster. In sectorization, groups should be very similar to each other, and therefore very different points can be grouped in the same sector. We classify sectorization problems into four types: basic sectorization, sectorization with service centers, resectorization, and dynamic sectorization. A Decision Support System for Sectorization, D3S, is being developed to deal with these four types of problems. Multi-objective genetic algorithms were implemented in D3S using Python, and a user-friendly web interface was developed using Django. Several applications can be solved with D3S, such as political districting, sales territory design, delivery service zones, and assignment of fire stations and health services to the population.

Abstract
In this paper, we investigate a new variant of the multi-commodity network flow problem, taking into consideration multiple transport lines and time windows. This variant arises in a city logistics environment, more specifically in a long-haul passenger transport system that is also used to transport urban freight. We propose two mixed integer programming models for two objective functions: minimization of network operational costs and minimization of travel times. Since the problems under study are NP-hard, we propose three size reduction heuristics. In order to assess the performance of the proposed algorithms, we carried out computational experiments on a set of synthetic problem instances. We use the relative percentage deviation as performance criterion. For the cost objective function, a LP-and-Fix algorithm outperforms other methods in most tested instances, but for the travel time, a hybrid method (size reduction with LP-and-Fix algorithm) is, in general, better than other approaches.

Abstract
Background: The multimedia capturing of live lectures has increased within higher education institutions, even in the pre-COVID-19 period. Despite student satisfaction, the video lecture capture (VLC) influence on students' attainment and achievement of intended learning outcomes is controversial. Methods: To explore the impact of VLC, a cross-sectional study across 2016/17 (n=209 students) and 2017/18 (n=206 students) was conducted in the course of Mechanistic Toxicology in Pharmaceutical Education. Results: The results showed that 73% and 90% of the assessed students entirely viewed the videos of theoretical (550 minutes) and practical/laboratory classes (250 minutes), respectively. VLC impacted student attainment and the achievement of intended learning outcomes on the capacity to understand the subjects and apply knowledge. Conclusion: The effectiveness of VLC is to be considered under the framework of constructive alignment and the specificities of the course.

Abstract
This study investigates the potential of dynamically adjusting WIP cap levels to maximize the throughput (TH) performance and minimize work in process (WIP), according to real-time system state arising from process variability associated with low volume and high-variety production systems. Using an innovative approach based on state-of-the-art deep reinforcement learning (proximal policy optimization algorithm), we attain WIP reductions of up to 50% and 30%, with practically no losses in throughput, against pure-push systems and the statistical throughput control method (STC), respectively. An exploratory study based on simulation experiments was performed to provide support to our research. The reinforcement learning agent's performance was shown to be robust to variability changes within the production systems.