Abstract Software-based artificial life methods use mathematical and computational models to mimic complexity in living systems. Although such methods seem promising pertaining to exploring emergent behaviour, obtaining the governing rulesets of such methods remains challenging. In this paper, we present a concept of combined use of methods targeting different levels/scales in an emergent behaviour to obtain software-based artificial life rulesets from observed data. Additionally, we investigate the consequences of using this combination of methods by proposing an instance of combining Cellular Automata and Agent-based modelling with Koopman-based linearization. Our experiments on systems of Elementary Cellular Automaton (Rule 30), Game of Life (GOL), and Vicsek’s flocking show that the combined method can learn the overall non-linear and emergent behaviour, and the underlying governing rulesets. Our research also indicates that by identifying several emergent scales or levels in a system, the combined method has the potential to shed light on the learnt system dynamics.

The scientific community has recently focused on intelligent models for predicting and optimizing EV energy management. Despite numerous studies in energy management optimization, there’s a critical need to address the trade-off between energy consumption and occupant comfort. Existing IoT systems face challenges in data analytics security and authenticity, highlighting the need for contemporary models to overcome data privacy and cost-related issues. This study introduces a smart contract model based on optimization and control modules, aiming to manage energy consumption while satisfying user comfort requirements intelligently. Introducing a smart contract model with hierarchical layers—prediction, optimization, control, and Blockchain—the proposed approach intelligently manages energy consumption while meeting user comfort requirements. Utilizing a Kalman filter for prediction and the BAT algorithm for optimization, the model integrates modules to tailor user preferences and enhance comfort. The synergy between the optimization module and a convolutional FLC enhances system performance, ensuring minimized energy usage and elevated user comfort levels. The study also evaluates the model’s implementation of the Hyperledger Fabric network, assessing outcomes regarding caliper, latency, throughput, and resource utilization.

In recent years, unmanned aerial vehicles (UAVs) have been broadly employed as a data collection platform to assist in efficiently collecting data from IoT devices. However, the deployment optimization of UAVs has been challenged due to the need to minimize the energy consumption of UAVs and IoT devices. Several algorithms have been recently proposed for tackling this challenge, but they still have room for improvement due to their slow convergence speed and memory-wasting problems. Therefore, in this study, a new energy-aware approach has been proposed for accurately optimizing the entire deployment of UAVs, which could minimize the total energy consumption. This approach is based on presenting a new encoding mechanism, namely an optimized population size mechanism, for representing both location and number of stop points in an effective manner. In this mechanism, similar to some studies in the literature, the whole population is responsible for the entire deployment, and each individual is responsible for a stop point in this deployment. However, this mechanism presents a novel way to optimize the number of stop points based on adding an auxiliary variable to each stop point to determine whether it will be removed, inserted, or replaced in the newly generated deployment. This variable will be optimized by the optimization techniques during the optimization process to search for the optimal choice for each stop point that could achieve a better deployment. Two well-known optimization techniques, known as differential evolution (DE) and gradient-based optimizer (GBO), are adapted using this mechanism to present new variants, namely DEoPS and GBoPS, for accurately tackling the deployment optimization problem. Two energy consumption formulations are used in our work to investigate the performance of DEoPS and GBoPS. Several experiments have been conducted to compare the performance of both DEoPS and GBoPS with several algorithms on eleven instances. The experimental findings show the effectiveness of GBoPS for the first formulation and the effectiveness of DEoPS for the second formulation.

Accurately estimating the unknown parameters of the photovoltaic (PV) models based on the measured voltage-current data is a challenging optimization problem due to its high nonlinearity and multimodality. An accurate solution to this problem is essential for efficiently simulating, controlling, and evaluating PV systems. There are three different PV models, including the single-diode model, the double-diode model, and the triple-diode model, with five, seven, and nine unknown parameters, respectively, proposed to represent the electrical characteristics of PV systems with varying levels of complexity and accuracy. In the literature, several deterministic and metaheuristic algorithms have been used to accurately solve this hard problem. However, due to the high nonlinearity of this problem, the deterministic methods could not achieve accurate solutions. On the other side, the metaheuristic algorithms, also known as gradient-free methods, could achieve somewhat good solutions for this problem, but they still need further improvements to strengthen their performance against stuck-in local optima and slow convergence speed problems. Over the last two years, several recent metaheuristic algorithms with better characteristics to improve convergence speed and avoid local optima have been proposed to tackle continuous optimization problems. However, the performance of the majority of those algorithms for estimating the parameters of PV models has not been investigated. Therefore, in this paper, the performance of nineteen recently published metaheuristic algorithms, such as the Mantis search algorithm (MSA), spider wasp optimizer (SWO), light spectrum optimizer (LSO), growth optimizer (GO), walrus optimization algorithm (WAOA), hippopotamus optimization algorithm (HOA), black-winged kite algorithm (BKA), quadratic interpolation optimization (QIO), sinh cosh optimizer (SCHA), exponential distribution optimizer (EDO), optical microscope algorithm (OMA), secretary bird optimization algorithm (SBOA), Parrot Optimizer (PO), Newton-Raphson-based optimizer (NRBO), crested porcupine optimizer (CPO), differentiated creative search (DCS), propagation search algorithm (PSA), one-to-one based optimizer (OOBO), and triangulation topology aggregation optimizer (TTAO), are studied to clarify their effectiveness in estimating the unknown parameters of PV models. In addition, those algorithms collaborate with two deterministic functions, namely the Lambert W-Function and the Newton-Raphson Method, to aid in solving the I-V curve equations more accurately, thereby improving the performance of PV systems. Those algorithms are assessed using four well-known PV solar cells and modules and compared with each other using several performance metrics, including best fitness, average fitness, worst fitness, standard deviation (SD), Friedman mean rank, and convergence speed; and a multiple-comparison test to compare the difference between their mean ranks. Results of this comparison show that SWO is more efficient and effective for SDM, DDM, and TDM over the majority of the studied PV solar cells and modules, and the Newton-Raphson Method is more efficient for solving the I-V curve equations. In addition, this study reports that the majority of the recently published metaheuristic algorithms perform poorly when applied to this problem.

Aging is inevitably associated with a decline in physical abilities and can pose challenges to the social lives of elderly individuals. In long-term care facilities, group exercise is instrumental for keeping elderly residents physically and socially healthy. Accommodating these needs in elderly care can be challenging due to staff shortages and other lacking resources. A robotic exercise coach could be helpful in such contexts. Intelligent human–robot interaction requires accurate and efficient human activity recognition. Several solutions focusing on human activity recognition in healthcare robotics have been proposed. However, multiperson activity recognition remains a challenging task in case of using vision-based or wearable sensors data, and past research has mainly focused on single-person rather than multiperson or group activity recognition. Moreover, the existing state-of-the-art methods for activity recognition mainly use heavyweight Convolutional Neural Network (CNN) models to achieve good accuracy. However, these models have certain drawbacks, such as requiring significant computational resources, higher memory and storage needs, and slower inference times. Another challenge is the limited number of publicly available datasets containing few activities for physical activity recognition. In this work, we propose a lightweight, deep learning-based, multiperson activity recognition system for group exercise training of elderly persons. Considering the limited publicly available datasets, we curated a new dataset named the Routine Exercise Dataset (RED), comprising 19 routine exercise activities recommended for elderly persons. The RED dataset has 14,440 samples collected from 19 participants and is one of the most extensive datasets of its kind. We evaluated our proposed activity recognition method based on proposed feature extraction modules and a one-dimensional multilayer long short-term memory network on 16 datasets, including 10 publicly available benchmark activity recognition datasets, an RED dataset, a publicly available dataset combined with RED dataset, and four noise-corrupted RED datasets. The results indicate the efficiency of the proposed method for real-time activity recognition compared to the state-of-the-art methods. The proposed method achieved F1-scores of 98.64%, 97.95%, and 99% on large-scale datasets named UESTC RGB-D, NTU RGB+D, and RED, respectively. We also developed a Robot Operating System (ROS)-based application to deploy our proposed system in a social robot and test it in real-life scenarios.

Guest Editorial: Special Issue on Generating Human Readable Explanations in NLP

E-waste stands for electronic devices, such as phones, computers, and televisions, that are disposed of when they reach the end of their useful lives. E-waste is becoming one of the most rapidly growing waste streams globally. The production of e-waste is rapidly increasing due to various factors, including the rapid advancement of technology, changing customer preferences, the widespread use of non-repairable parts, and deliberate planned obsolescence during the design of such products to encourage consumers to replace their products more often, boosting sales and profits. E-waste contains precious and rare metals such as gold, silver, copper, indium, and palladium. When these materials are not properly recovered, recycled, and reused, the production of new electronics will require the mining of finite natural resources, leading to environmental damage and the depletion of resources. As of early 2019, Norway has established an Extended Producer Responsibility (EPR) system, which requires producers and importers to finance and ensure proper collection and recycling of end-of-life (EoL) e-waste, resulting in an increase in the e-waste recycling rate to approximately 91% in Norway. However, electronic waste such as mobile phones and computers often goes unrecycled. Currently, 82% of Norwegian households have at least one extra mobile phone that they are not using. People may hesitate to dispose of their electronic waste and tend to stockpile it at home for several reasons. These include feeling emotionally attached to the item due to personal memories, not knowing where or how to dispose of it, being unaware of the potential environmental and health hazards associated with improper disposal and having concerns about the data security and personal information stored on the device. To address these issues, the paper introduces a cutting-edge tracking and tracing platform that features an intuitive user interface for both users and administrators. With the proposed platform, users and administrators can easily access essential information regarding their e-waste disposal. This convenient and efficient system offers a practical solution for tracking e-waste, thereby promoting trust between device owners and the disposal process. Furthermore, the platform incentivizes users by providing them with valuable information regarding their e-waste disposal, creating a positive impact on environmental sustainability.

An evaluation is administered to measure students’ learning outcomes, which nowadays become challenging for instructors as student growth increases exponentially. Several models are proposed in the literature based on selected artificial intelligence algorithms that are once trained and then deployed. The problem with these kinds of systems is that the trained models are locked and cannot adjust to dynamically changing circumstances, leading to a drop in performance. Moreover, these systems only considered basic parameters for computing the semantic similarity, resulting in less accuracy. This paper develops an intelligent student evaluation model based on a predictive optimization approach, which considers question type, structure, necessary keywords, language, and conceptual aspects to evaluate the student’s answer. In order to enhance the performance of the proposed evaluation system, we have proposed a predictive optimization approach where a deep neural network is used as a learning module to learn from training data, and particle swarm optimization and gradient descent are used as an optimization scheme to optimize weighting parameters for the deep neural network. The proposed work uses and analyzes the real dataset of NTNU students’ exams to validate the proposed platform’s practicability. Initially, we will employ the natural language processing technique of deep learning in which semantic similarity score and other features will be used to compute the degree of relevance between actual answers and students’ provided answers. The proposed semantic similarity score algorithm is based on the WordNet library and Growbag dataset to check the solution’s semantics, conceptual aspects, and creativity. The resulting score will be used as a supervised machine-learning classification system feature. Performance of the classification model will be ensured using standard evaluation measures, including Precision, recall, and f-measure. The end goal of this platform is to acquire the grade against the student’s answer given as input in the developed platform.

Efficient energy demand management plays an essential role in smart grid, sustainable and smart cities applications and efforts to reduce CO2 emissions. In this paper, we propose a framework for describing the household daily energy consumption and how it can be used to help residential households to perform appliance rescheduling to reduce energy consumption and hence reducing their energy bills while keeping resident’s comfort. In this paper, heuristic optimization techniques such as genetic algorithm (GA) and particle swarm optimization (PSO) are used for solving the load scheduling problem. Due to its ability to deal with computational complex scenarios in less computational time using less and less computational resources, Heuristic optimization techniques are used. In the proposed model, dynamic pricing is adopted where the objective is to minimize the overall cost of electricity consumption and payments by scheduling different devices in a way that fulfil each individual’s constraints and preferences. Here, MATLAB was used as the simulation platform. Simulation results showed that GA and PSO can optimize energy consumption and bills and at the same time fulfils needs and preferences of each individual customer.

Light modelling, simulation, and photometric calculations are by now common tasks in the lighting design process. These practices contribute to the definition and comparison of suitable layout arrangements and help predict the impact of lighting devices. Those tasks demand the use of tools to support the simulation of different scenarios, the analyses of their pros and cons according to different criteria (e.g., health and safety, perception, aesthetics, energy consumption, and costs), and decision-making. Digital twins have emerged as relevant technologies to simulate and visualize different ``what-if'' scenarios associated with physical entities and processes. In this paper, we investigate the state-of-the-art research concerning the use of digital twins for supporting lighting analysis in the urban/outdoor context. We also present and discuss challenges and research opportunities related to the design, implementation, and validation of digital twins in this domain.

Abstract In today’s world, lung cancer is a significant health burden, and it is one of the most leading causes of death. A leading type of lung cancer is malignant mesothelioma (MM). Most of the MM patients do not show any symptoms. Etiology plays a vital factor in the diagnosis of any disease. Positron emission tomography (PET), magnetic resonance imaging (MRI), biopsies, X-rays and blood tests are essential but costly and invasive MM risk factor identification methods. In this work, we mainly focused on the exploration of the MM risk factors. The identification of mesothelioma symptoms was carried out by utilizing the data of mesothelioma patients. However, the dataset was comprised of both healthy and mesothelioma patients. The dataset is prone to a class imbalance problem in which the number of MM patients significantly less than healthy individuals. To overcome the class imbalance problem, the synthetic minority oversampling technique has been utilized. The association rule mining-based Apriori algorithm has been applied to a preprocessed dataset. Before using the Apriori algorithm, both duplicate and irrelevant attributes were removed. Moreover, the numerical attributes were also classified into nominal attributes and the association rules were generated in the dataset. Our results show that erythrocyte sedimentation rate, asbestos exposure and its duration time, and pleural and serum lactic dehydrogenase ratio are major risk factors of MM. The severe stages of MM can be avoided by earlier identification of risk factors of the disease. The failure of identification of risk factors can lead to increased risk of multiple medical conditions, including cardiovascular diseases, mental distress, diabetes and anemia.

In a wireless sensor network (WSN), node localization is a key requirement for many applications. The concept of mobile anchor-based localization is not a new concept; however, the localization of mobile anchor nodes gains much attention with the advancement in the Internet of Things (IoT) and electronic industry. In this paper, we present a range-free localization algorithm for sensors in a three-dimensional (3D) wireless sensor networks based on flying anchors. The nature of the algorithm is also suitable for vehicle localization as we are using the setup much similar to vehicle-to-infrastructure- (V2I-) based positioning algorithm. A multilayer C-shaped trajectory is chosen for the random walk of mobile anchor nodes equipped with a Global Positioning System (GPS) and broadcasts its location information over the sensing space. The mobile anchor nodes keep transmitting the beacon along with their position information to unknown nodes and select three further anchor nodes to form a triangle. The distance is then computed by the link quality induction against each anchor node that uses the centroid-based formula to compute the localization error. The simulation shows that the average localization error of our proposed system is 1.4 m with a standard deviation of 1.21 m. The geometrical computation of localization eliminated the use of extra hardware that avoids any direct communication between the sensors and is applicable for all types of network topologies.

Text representation has a critical impact on the accuracy of text classifiers which is imperative to be strengthened. On the other hand, the question of how the state-of-the-art embeddings outperform previous approaches cannot be well explained. To advance text representation and better understand the internal mechanism, we propose a novel end-to-end framework named Ensemble Framework for Text Embedding (EFTE), which weightedly combines diverse embeddings and simultaneously represents sentences' and tokens' features in a more reasonable way. According to the experimental results in sentiment classification, our proposed embedding apparently improves the effectiveness compared to six single embeddings. Moreover, the importance of each embedding in terms of EFTE integration and how different embeddings influence the results by classification are discussed.

Accurate prediction of soil erosion rate is a quite important issue for a wise and sustainable use of soil resources. In this study, an adaptive neuro-fuzzy inference system (ANFIS) approach is used to construct a prediction model. The objectives of this study is to develop fuzzy logic models that predict soil erosion in a relatively large watershed using a limited number of input variables, compare the predictions of soil erosion using ANFIS model with those of the Revised Universal Soil Loss Equation RUSLE. With the incorporation of Geographical Information System (GIS), it is possible to analyse satellite data, which gives required information like land use and cover, slope, distribution of rainfall, flow direction etc. of study watershed. The capabilities of these technologies increase when they are integrated with ANFIS model for erosion prediction. ANFIS model and GIS integrated erosion prediction models do not only estimate soil loss but also provide the spatial distributions of the erosion. Generating accurate erosion risk maps in GIS environment is very important to locate the areas with high erosion risks for prioritization and to develop adequate conservation techniques for a better sustainable management of Wadi Sahel watershed (Algeria).

In this paper we present a recently developed intelligent winch prototyping tool for optimising the design of maritime winches, continuing our recent line of work using artificial intelligence for intelligent computer-automated design of offshore cranes. The tool consists of three main components: (i) a winch calculator for determining key performance indicators for a given winch design; (ii) a genetic algorithm that interrogates the winch calculator to optimise a chosen set of design parameters; and (iii) a web graphical user interface connected with (i) and (ii) such that winch designers can use it to manually design new winches or optimise the design by the click of a button. We demonstrate the feasibility of our work by a case study in which we improve the torque profiles of a default winch design by means of optimisation. Extending our generic and modular software framework for intelligent product optimisation, the winch calculator can easily be interfaced to external product optimisation clients by means of the HTTP and WebSocket protocols and a standardised JSON data format. In an accompanying paper submitted concurrently to this conference, we present one such client developed in Matlab that incorporates a variety of intelligent algorithms for the optimisation of maritime winch design.

In this study, the effect of concentration, intensification and dilation of three common linguistic hedges (LHs), namely, very, indeed, and more or less on the performance of a fuzzy system for evaluating student's academic evaluation is presented. A LH may be viewed as an operator that acts on a fuzzy set representing the meaning of its operand. As an example, the operator very acts on the fuzzy meaning of the term high grade to have a secondary meaning of very high grade. This property changes the shape of the fuzzy sets and hence the amount of overlap between adjacent sets. It, in turn, improves the meaning of the fuzzy rules and hence the accuracy of the proposed fuzzy evaluation systems. The proposed LHs based fuzzy evaluator systems are compared with a standard fuzzy sets based fuzzy evaluator system using an example drawn from literature. Empirical results of the example presented in this paper show that concentration and dilation effect of LHs is not significant compared to standard fuzzy sets.

Examination is one of the common assessment methods to assess the level of knowledge of students. Assessment methods probably have a greater influence on how and what students learn than any other factor. Assessment is used to discriminate not only between different students but also between different levels of thinking. Due to the increasing trends in class sizes and limited resources for teaching, the need arises for exploring other assessment methods. Multiple-Choice Questions (MCQs) have been highlighted as the main way of coping with the large group teaching, ease of use, testing large number of students on a wide range of course material, in a short time and with low grading costs. MCQs have been criticised for encouraging surface learning and its unfairness. MCQs have a variety of scoring options; the most widely used method is to compute the score by only focusing on the responses that the student made. In this case, the number of correct responses is counted, the number of incorrect answers is counted and a final score is reported as either the number of the correct answers or the number of correct answers minus the number of incorrect answers. The disadvantages of this approach are that other dimensions such as importance and complexity of questions are not considered, and in addition, it cannot discriminate between students with equal total score. In this paper, a method to automatically evaluate MCQs considering importance and complexity of each question and providing a fairer way to discriminating between students with equal total scores is presented.

The proper system for evaluating the learning achievement of students is the key to realizing the purpose of education and learning. Traditional grading methods are largely based on human judgments, which tend to be subjective. In addition, it is based on sharp criteria instead of fuzzy criteria and suffers from erroneous scores assigned by indifferent or inexperienced examiners, which represent a rich source of uncertainties, which might impair the credibility of the system. In an attempt to reduce uncertainties and provide more objective, reliable, and precise grading, a sophisticated assessment approach based on type-2 fuzzy set theory is developed. In this paper, interval type-2 (IT2) fuzzy sets, which are a special case of the general T2 fuzzy sets, are used. The transparency and capabilities of type-2 fuzzy sets in handling uncertainties is expected to provide an evaluation system able to justify and raise the quality and consistency of assessment judgments.

For many years, NTNU in Ålesund (formerly Aalesund University College) has maintained a close relationship with the maritime industrial cluster, centred in the surrounding geographical region, thus acting as a hub for both education, research, and innovation. Of many common relevant research topics, virtual prototyping is currently one of the most important. In this paper, we describe our first complete version of a generic and modular software framework for intelligent computer-automated product design. We present our framework in the context of design of offshore cranes, with easy extensions to other products, be it maritime or not. Funded by the Research Council of Norway and its Programme for Regional R&D and Innovation (VRI), the work we present has been part of two separate but related research projects (grant nos. 241238 and 249171) in close cooperation with two local maritime industrial partners. We have implemented several software modules that together constitute the framework, of which the most important are a server-side crane prototyping tool (CPT), a client-side web graphical user interface (GUI), and a client-side artificial intelligence for product optimisation (AIPO) module that uses a genetic algorithm (GA) library for optimising design parameters to achieve a crane design with desired performance. Communication between clients and server is achieved by means of the HTTP and WebSocket protocols and JSON as the data format. To demonstrate the feasibility of the fully functioning complete system, we present a case study where our computer-automated design was able to improve the existing design of a real and delivered 50-tonnes, 2.9 million EUR knuckleboom crane with respect to some chosen desired design criteria. Our framework being generic and modular, both clientside and server-side modules can easily be extended or replaced. We demonstrate the feasibility of this concept in an accompanying paper submitted concurrently, in which we create a simple product optimisation client in Matlab that uses readily available toolboxes to connect to the CPT and optimise various crane designs by means of a GA. In addition, our research team is currently developing a winch prototyping tool to which our existing AIPO module can connect and optimise winch designs with only small configuration changes. This work will be published in the near future.

In an accompanying paper submitted concurrently to this conference, we present our first complete version of a generic and modular software framework for intelligent computer-automated product design. The framework has been implemented with a client-server software architecture that automates the design of offshore cranes. The framework was demonstrated by means of a case study where we used a genetic algorithm (GA) to optimise the crane design of a real and delivered knuckleboom crane. For the chosen objective function, the optimised crane design outperformed the real crane. In this paper, we augment our aforementioned case study by implementing a new crane optimisation client in Matlab that uses a GA both for optimising a set of objective functions and for multi-objective optimisation. Communicating with an online crane prototyping tool, the optimisation client and its GA are able to optimise crane designs with respect to two selected design criteria: the maximum safe working load and the total crane weight. Our work demonstrates the modularity of the software framework as well as the viability of our approach for intelligent computer-automated design, whilst the results are valuable for informing future directions of our research.

Assessment and grading practices have the potential not only to measure and report learning but also to promote it. Student assessment is integral to learning experience and to curriculum design. It is typically used to evaluate learning outcomes and provide the basis for certification of individual students. Conventional grading system is largely based on human judgments, which tend to be subjective and have high degrees of errors and uncertainties. Because of the traditional pressure in assessment towards objectivity, conformity, consistency and certainty and due to the increasing trends in class sizes and limited resources for teaching, examiners and lectures are always challenging their abilities to deliver timely and fair assessments. To cope with the aforementioned challenges, the need arises for exploring innovation and technology to facilitate assessment and to incorporate other dimensions, which could not be considered in conventional grading system, to ensure and promote deep learning and critical thinking. In this paper a fuzzy grading system, which considers complexity, difficulty and importance of exam questions, is presented. A simplified implementation of interval type-2 fuzzy system using the basic knowledge of type-1 fuzzy is presented. A comparison between the use of type-1 and interval type-2 fuzzy systems in reducing uncertainties and providing more transparent and fair assessment that can reflect needs of individual students and foster development is presented.

Conventional agricultural fields are sprayed uniformly to control weeds, insects, and diseases. To reduce cultivation expenses, to produce healthier food and to create more environmentally friendly farms, chemicals should only be applied to the right place at the right time and exactly with the right amount. In this article, a task and motion planning for a team of autonomous vehicles to reduce chemicals in farming is presented. Field data are collected by small unmanned helicopters equipped with a range of sensors, including multispectral and thermal cameras. Data collected are transmitted to a ground station to be analyzed and triggers aerial and ground-based vehicles to start close inspection and/or plant/weed treatment in specified areas. A complete trajectory is generated to enable ground-based vehicle to visit infested areas and start chemical/mechanical weed treatment.