Class (philosophy)

About: Class (philosophy) is a research topic. Over the lifetime, 821 publications have been published within this topic receiving 28000 citations.

Deep Policy Dynamic Programming for Vehicle Routing Problems

Abstract: AbstractRouting problems are a class of combinatorial problems with many practical applications. Recently, end-to-end deep learning methods have been proposed to learn approximate solution heuristics for such problems. In contrast, classical dynamic programming (DP) algorithms guarantee optimal solutions, but scale badly with the problem size. We propose Deep Policy Dynamic Programming (DPDP), which aims to combine the strengths of learned neural heuristics with those of DP algorithms. DPDP prioritizes and restricts the DP state space using a policy derived from a deep neural network, which is trained to predict edges from example solutions. We evaluate our framework on the travelling salesman problem (TSP), the vehicle routing problem (VRP) and TSP with time windows (TSPTW) and show that the neural policy improves the performance of (restricted) DP algorithms, making them competitive to strong alternatives such as LKH, while also outperforming most other ‘neural approaches’ for solving TSPs, VRPs and TSPTWs with 100 nodes.KeywordsDynamic ProgrammingDeep LearningVehicle Routing

A modified extragradient algorithm for a certain class of split pseudo-monotone variational inequality problem

Abstract: In this paper, we introduce and study a modified extragradient algorithm for approximating solutions of a certain class of split pseudo-monotone variational inequality problem in real Hilbert spaces. Using our proposed algorithm, we established a strong convergent result for approximating solutions of the aforementioned problem. Our strong convergent result is obtained without prior knowledge of the Lipschitz constant of the pseudo-monotone operator used in this paper, and with minimized number of projections per iteration compared to other results on split variational inequality problem in the literature. Furthermore, numerical examples are given to show the performance and advantage of our method as well as comparing it with related methods in the literature.

Context-oriented programming

Abstract: Context-oriented programming (COP) introduces one more notion to reason about the structure of software systems: a context (an environment) is a set of entities bound with a system of relations. This view is applicable where the object-oriented one is inadequate. Implementation of COP requires the same techniques as OOP: COP and OOP are different things assembled from the same components. COP allows things that OOP cannot do, for example, COP enables us to use late binding for elementary data that are not OOP objects. 1. The purpose of the work The key element of software activities (like designing, writing, debugging, reading and modifying programs) is reasoning. We do not give a definition for reasoning, but we assume that the use of a small number of adequate concepts facilitates reasoning—that is, if the concepts in use are adequate and expressive, if no auxiliary concepts are required, it is easier to reason about programs. The purpose of this work is to extend the set of notions, which may be used to reason about the problem to be solved and the structure of the corresponding software system. Along with the notion of a context, an approach to problem decomposition, and a programming technique appear. 2. The method of COP 2.1. Quick introduction An environment (a context) expresses the idea of a system of relations (where "relation" is used in its everyday meaning, a connexion between things). The phrase "system of relations" implies that there are some entities, upon which relations are established. The key point in COP is that to use the relations, we do not need to know details about the entities. This is good, but how can we use the relations? In principle, different approaches are possible. The simplest way is to implement a set of functions, which imply that the data elements to be processed are bound with these relations. (Indeed, using these functions, we use knowledge of that the data elements are related, otherwise it would be incorect to use the functions). The next step is to introduce polymorphism. This may be done via tables of pointers to functions, that is, virtual method tables (VMT, where "method" is understood as "member function"). Contexts are implemented as VMT's, and these VMT's are not bound to data. So, from the implementation point of view, there are interchangeable sets of functions, allowing us to work with different (but somehow similar) sets of objects; the objects in a set do not necessarily form a datastructure, and may have different lifetimes. It should be also noted that COP introduces two levels at which we can work with contexts. At the level of implementation, we know all details about related entities that constitute the context. At the level of use, we do not have full access to the entities; we can use only a set of functions allowing us to work with the entities, but this should be enough to accomplish our tasks. In exchange, this restricted access allows us to work in different contexts (of the same "class") in a uniform way.

Class DojoLMS in the Interactive Learning of PAUD Educators in the Disruption Era 4.0

Abstract: With the change in the learning system from conventional to online in the current pandemic, it is a challenge for teachers, especially early childhood educators, to find interactive learning methods. Therefore, in this paper, we will focus on the Learning Management System (LMS) strategy used by educators to continue to provide quality interactive learning to early childhood education level students. Because until now, the problem faced by Early Child Education (ECE) teachers is that there is no research on dual-function learning applications, both for communication and giving learning assignments. The research will be conducted using the slovin method and surveys of 70 teacher respondents who have experience using learning applications and 30 of them are experts in assessing the visualization of a system. interviews were conducted with a time span of one month from February - March 2021. Aim of the research is to provide learning application solutions that have dual functions and are interactive for early childhood targets. An Interactive learning application for the main target of early childhood, where from the learning media studied ranging from Class Dojo, Whatsapp, Google meet, Youtube, Worksheets, and email, Class dojo which received high votes from 70 respondents by getting 72% results became findings study. This dojo class application can be a recommendation for educators' learning media because it has complete features and can be improved communication between educators and guardians of students through the additional contact feature of the student guardian group of the dojo application.