School-university cluster of disciplines developing the calculative-algorithmic component of computational thinking

Problem relevance. Due to the multi-departmental concepts and the different content of educational programs of schools and universities, a serious problem arises of the succession and continuity of the education system along the “vertical” in general and subject teaching in particular. Another didactic problem is the need to ensure interdisciplinary connections of basic courses in the traditional disciplinary model of the educational process for more effective and expedient formation of certain student’s competencies sets. In this regard, it is of interest to create new organizational and meaningful approaches to training specialists without a significant restructuring of the traditional educational process.The purpose of the article is to substantiate a collaborative model of subject training of students in a school-university cluster of disciplines, which ensures the succession and continuity of education at school and university.Methodological basis. On the example of three disciplines “Programming”, “Computanional Methods”, “Information Technologies in Education”, a cluster model of teaching schoolchildren and students in the direction of training “Mathematics and Computer Science” has been designed and implemented. A feature of the considered school-university cluster of disciplines is a unified methodological base of target, meaningful and didactic elements that form and develop the calculative-algorithmic component of the computational thinking of students. The basis of the means and methods of teaching in the cluster is made up of cognitive techniques and a platform of “computational and algorithmic primitives” — solving elementary task template. A recursive approach is used in the methods of cluster subject teaching of schoolchildren and students.Results and Conclusions. The model of the created disciplinary cluster “Programming — Computanional Methods — Information Technologies in Education” contributes to the formation and development of the calculative-algorithmic component of the computational thinking of schoolchildren and students, and also forms their assigned groups of competencies. The school-university cluster of disciplines ensures real succession and continuity of school and university education, without unnecessary, sometimes artificial, labor-intensive additional organizational and methodological means and techniques. The approach under consideration can be used to create clusters of disciplines in various educational areas, allowing their meaningful collaboration and forming given competencies sets and schoolchildren’s and student’s cognitive abilities.

Algorithmic Component of the New Statistical Quality Control Software Module

Great development of information technologies largely determines the content of modern management, providing managers abilities for automatical data collection and processing, making decisions based on the use of a wide range of application software for various purposes. The use of quality management methods and tools is an important condition for product competitiveness. The possibility of using the software module as part of the developed information system is proposed. The article gprovides detailed description of statistical software module development for the purpose of product quality control. It is concluded that it is expedient and necessary to introduce this development into production.

Infeasibility and Structural Bias in Differential Evolution

This paper investigates a range of popular differential evolution (DE) configurations to identify components responsible for emergence of structural bias – a recently identified tendency of algorithms to prefer some regions of search space over others, for reasons unrelated to objective function values. Previous work has explored this tendency for genetic algorithms (GA) and particle swarm optimisation (PSO), finding a relationship between population size and extent of structural bias, hence highlighting potential weaknesses of those algorithms. In current article, we focus on DE, extend the investigation to include consideration of an algorithmic component that is often overlooked – constraint handling mechanism. Towards this end, a wide range of DE configurations was tested here. Results suggest that DE is generally robust to structural bias. Unlike the case with GA and PSO, population size seems to have no influence on DE structural bias. Only one of variants studied – DE/current-to-best/1/bin – shows clear signs of bias, however, we show that this effect is mitigated by a judicious choice of constraint handling technique. These findings contribute towards explaining widespread success of DE variants in algorithm comparison studies; its robustness to structural bias represents the absence of a factor that may confound other algorithms.

Evolutionary online behaviour learning and adaptation in real robots

Online evolution of behavioural control on real robots is an open-ended approach to autonomous learning and adaptation: robots have the potential to automatically learn new tasks and to adapt to changes in environmental conditions, or to failures in sensors and/or actuators. However, studies have so far almost exclusively been carried out in simulation because evolution in real hardware has required several days or weeks to produce capable robots. In this article, we successfully evolve neural network-based controllers in real robotic hardware to solve two single-robot tasks and one collective robotics task. Controllers are evolved either from random solutions or from solutions pre-evolved in simulation. In all cases, capable solutions are found in a timely manner (1 h or less). Results show that more accurate simulations may lead to higher-performing controllers, and that completing the optimization process in real robots is meaningful, even if solutions found in simulation differ from solutions in reality. We furthermore demonstrate for the first time the adaptive capabilities of online evolution in real robotic hardware, including robots able to overcome faults injected in the motors of multiple units simultaneously, and to modify their behaviour in response to changes in the task requirements. We conclude by assessing the contribution of each algorithmic component on the performance of the underlying evolutionary algorithm.

odNEAT: An Algorithm for Decentralised Online Evolution of Robotic Controllers

Online evolution gives robots the capacity to learn new tasks and to adapt to changing environmental conditions during task execution. Previous approaches to online evolution of neural controllers are typically limited to the optimisation of weights in networks with a prespecified, fixed topology. In this article, we propose a novel approach to online learning in groups of autonomous robots called odNEAT. odNEAT is a distributed and decentralised neuroevolution algorithm that evolves both weights and network topology. We demonstrate odNEAT in three multirobot tasks: aggregation, integrated navigation and obstacle avoidance, and phototaxis. Results show that odNEAT approximates the performance of rtNEAT, an efficient centralised method, and outperforms IM-([Formula: see text]), a decentralised neuroevolution algorithm. Compared with rtNEAT and IM-([Formula: see text]), odNEAT’s evolutionary dynamics lead to the synthesis of less complex neural controllers with superior generalisation capabilities. We show that robots executing odNEAT can display a high degree of fault tolerance as they are able to adapt and learn new behaviours in the presence of faults. We conclude with a series of ablation studies to analyse the impact of each algorithmic component on performance.

Interactive Policy Learning through Confidence-Based Autonomy

We present Confidence-Based Autonomy (CBA), an interactive algorithm for policy learning from demonstration. The CBA algorithm consists of two components which take advantage of the complimentary abilities of humans and computer agents. The first component, Confident Execution, enables the agent to identify states in which demonstration is required, to request a demonstration from the human teacher and to learn a policy based on the acquired data. The algorithm selects demonstrations based on a measure of action selection confidence, and our results show that using Confident Execution the agent requires fewer demonstrations to learn the policy than when demonstrations are selected by a human teacher. The second algorithmic component, Corrective Demonstration, enables the teacher to correct any mistakes made by the agent through additional demonstrations in order to improve the policy and future task performance. CBA and its individual components are compared and evaluated in a complex simulated driving domain. The complete CBA algorithm results in the best overall learning performance, successfully reproducing the behavior of the teacher while balancing the tradeoff between number of demonstrations and number of incorrect actions during learning.