Recommender systems (RSs) are software systems that use machine learning techniques to suggest items, such as movies, products or routes to users based on input provided over time. These systems have been widely used by companies, such as Amazon, Google, and Netflix, who rely heavily on recommendations as a core element of their interaction with users. In the case of software engineering, RSs are used to recommend, for example, software tasks, developers, APIs, libraries, and bug fixes. More recently, advancements in natural language processing and large language models (LLMs) have led to the creation of LLM agents capable of generating plans, interacting with users and the environment, receiving feedback, and using tools. These developments have resulted in the use of LLM agents in RSs, giving rise to multi-agent RSs referred to as agentic recommender systems (ARSs). However, agentic RSs are not yet well characterized in terms of the basic elements of system modeling and design. This systematic literature review characterizes agentic recommender systems with respect to agents, roles, relationships, prompts, integration, use cases, strategies and evaluation methods. An analysis of published studies yields insights, such as (i) the identification of 13 distinct types of agents, (ii) the popularity of GPT models, (iii) a framework for ARSs, (iv) a generalized prompt, (v) the use of persona, cues, output format, and zero-shot prompting, (vi) the frequent use of Amazon, Yelp, and MovieLens datasets, (vii) the use of nDCG@k and Recall@k as common evaluation metrics, and (viii) the identification of many research gaps. This systematic literature review aims to support the continued research and development of agentic recommender systems.

Depression has impacted millions of people world-wide and has become one of the most prevalent mental disorders. Early mental disorder detection can lead to cost savings for public health agencies and avoid the onset of other major comorbidities. Additionally, the shortage of specialized personnel is a critical issue because clinical depression diagnosis is highly dependent on expert professionals and is time-consuming.In this study, we explore the use of GPT-4 for clinical depression assessment based on transcript analysis. We examine the model’s ability to classify patient interviews into binary categories: depressed and not depressed. A comparative analysis is conducted considering prompt complexity (e.g., using both simple and complex prompts), as well as varied temperature settings, to assess the impact of prompt complexity and randomness on the model’s performance. Results indicate that GPT-4 exhibits considerable variability in accuracy and F1-Score across configurations, with optimal performance observed at lower temperature values (0.0-0.2) for complex prompts. However, beyond a certain threshold (temperature ≥ 0.3), the relationship between randomness and performance becomes unpredictable, diminishing the gains from prompt complexity. These findings suggest that, while GPT-4 shows promise for clinical assessment, the configuration of the prompts and model parameters requires careful calibration to ensure consistent results. This preliminary study contributes to understanding the dynamics between prompt engineering and large language models, offering insights for future development of AI-powered tools in clinical settings.

Agentic AI describes the use of LLMs in novel AI agents that can answer questions or collaborate to achieve goals. These LLM agents can be used to build a novel generation of recommender systems. However, little is known about the LLM agents or their relationships needed to provide recommendations. Once identified, a framework can be constructed. Moreover, evaluating this framework is still not well understood. In this paper, we propose an agentic AI-based, multi-agent framework for recommender systems. We first identify LLM agents proposed in the literature, followed by the identification of their relationships and we propose a framework to represent them. Next, we evaluate this framework with respect to the LLM agents and functionalities of a recommender system based on published studies. This study is a stepping stone in a novel paradigm shift in the construction of recommender systems.

Business processes are commonly represented by modelling languages, such as Event-driven Process Chain (EPC), Yet Another Workflow Language (YAWL), and the most popular standard notation for modelling business processes, the Business Process Model and Notation (BPMN). Most recently, chatbots, programs that allow users to interact with a machine using natural language, have been increasingly used for business process execution support. A recent category of chatbots worth mentioning is generative-based chatbots, powered by Large Language Models (LLMs) such as OpenAI’s Generative Pre-Trained Transformer (GPT) model and Google’s Pathways Language Model (PaLM), which are trained on billions of parameters and support conversational intelligence. However, it is not clear whether generative-based chatbots are able to understand and meet the requirements of constructs such as those provided by BPMN for process execution support. This paper presents a case study to compare the performance of prominent generative models, GPT and PaLM, in the context of process execution support. The research sheds light into the challenging problem of using conversational approaches supported by generative chatbots as a means to understand process-aware modelling notations and support users to execute their tasks.

This paper introduces the ‘GPT-in-the-loop’ approach, which seeks to investigate the reasoning capabilities of Large Language Models (LLMs) like Generative Pre-trained Transformers (GPT) within multiagent systems (MAS). Moving beyond traditional adaptive approaches that generally require long training processes, our framework employs GPT-4 to enhance problem-solving and explanation skills. To explore this approach, we apply it to a smart streetlight application in the Internet of Things (IoT) context, wherein each streetlight is controlled by an autonomous agent equipped with sensors and actuators, tasked with creating an energy-efficient lighting system. With the integration of GPT-4, these agents have shown enhanced decision-making and adaptability, without necessitating prolonged training. We compare this approach with both conventional neuroevolutionary methods and manually crafted solutions by software engineers, underscoring the potential of GPT-driven behavior in multiagent systems. It is important to note that these comparisons are preliminary, and further, more extensive testing is critical to determine the approach’s applicability across a wider range of MAS scenarios. Structurally, the paper delineates the incorporation of GPT into the agent-driven Framework for the Internet of Things (FIoT), details our proposed GPT-in-the-loop approach, presents comparative results within the IoT setting, and concludes with insights and prospective future directions.

There is an increasing interest in leveraging Large Language Models (LLMs) for managing structured data and enhancing data science processes. Despite the potential benefits, this integration poses significant questions regarding their reliability and decision-making methodologies. Our objective is to elucidate and express the factors and assumptions guiding GPT-4’s model selection recommendations. It highlights the importance of various factors in the model selection process, including the nature of the data, problem type, performance metrics, computational resources, interpretability vs accuracy, assumptions about data, and ethical considerations. We employ a variability model to depict these factors and use toy datasets to evaluate both the model and the implementation of the identified heuristics. By contrasting these outcomes with heuristics from other platforms, our aim is to determine the effectiveness and distinctiveness of GPT-4’s methodology. This research is committed to advancing our comprehension of AI decision-making processes, especially in the realm of model selection within data science. Our efforts are directed towards creating AI systems that are more transparent and comprehensible, contributing to a more responsible and efficient practice in data science.

Identifying regions of high demand for transportation services can help drivers maximize their profits, assist companies in dynamic pricing or resource allocation, and reduce passenger’s wait times. However, their identification is not trivial because of the many factors that impact the demand, such as the weather or time of the day, and the possible lack of communication between drivers. In this paper, we present a framework to identify regions of high demand for transportation services based on cluster evolution and graph analysis. The framework identifies how clusters of moving objects evolve, creates a graph to represent the evolution, and use cluster relationships to calculate a rate of change based on the objects that enter of leave the cluster. Results can be described based on the day or the evolution of a cluster. A use case with taxis in Rome is performed and two main regions of high demand for taxis are identified, one near hotels and another near a bus, taxi, and train terminal.

Graphs have been used as the foundation for several types of analysis methods in multiple application domains. In our study, we investigate spatial-temporal data analysis using graphs that capture knowledge about clusters of moving objects, their relationships, and their evolution. The study cope with the need to provide analysis techniques that consider cluster evolution. In this paper, we describe our ongoing study on knowledge graphs in spatial-temporal cluster evolution and potential research directions.

Data science projects often involve various machine learning (ML) methods that depend on data, code, and models. One of the key activities in these projects is the selection of a model or algorithm that is appropriate for the data analysis at hand. ML model selection depends on several factors, which include data-related attributes such as sample size, functional requirements such as the prediction algorithm type, and nonfunctional requirements such as performance and bias. However, the factors that influence such selection are often not well understood and explicitly represented. This paper describes ongoing work on extending an adaptive variability-aware model selection method with bias detection in ML projects. The method involves: (i) modeling the variability of the factors that affect model selection using feature models based on heuristics proposed in the literature; (ii) instantiating our variability model with added features related to bias (e.g., bias-related metrics); and (iii) conducting experiments that illustrate the method in a specific case study to illustrate our approach based on a heart failure prediction project. The proposed approach aims to advance the state of the art by making explicit factors that influence model selection, particularly those related to bias, as well as their interactions. The provided representations can transform model selection in ML projects into a non ad hoc, adaptive, and explainable process.

Smart wearable devices for patient monitoring rely on batteries as energy-source for capturing vital signs, processing information locally, and transmitting data. The advantages of such solutions are providing mobility to users, connectivity to send data constantly, and low cost. These devices are wireless and must be tiny to be carried comfortably by the users. This fact restricts energy autonomy and requires frequent replacement or recharge of batteries. The highest energy cost is commonly attributed to transmissions in wireless devices, and several studies focused on communication and routing protocols to enhance energy efficiency in such solutions. However, researchers should give more attention to data acquisition of physiological sensors regarding energy efficiency in such solutions. In this preliminary study, we present the effects of a self-adaptive algorithm on the energy consumption of popular wearable physiological sensors. Our prototype is composed of an oximeter and a temperature sensor. Our experiments demonstrate that the self-adaptive procedure can save up to 80% energy consumption regarding the oximeter when monitoring stable patients at low risk and 51% in unstable patients. In addition, the temperature sensor can reach 97% of energy savings in the self-adaptive mode. The sensors’ data acquisition can present a superior energy cost than radio transmissions on such devices. In future work, we will explore the potential benefits of the algorithm in all main activities of our monitoring device.

Wearable sensors utilize small, low-cost, noninvasive, and wireless components. These sensors capture vital signs, allowing the monitoring of patients remotely. In this manner, they are efficient tools to enhance patient care and can be used to monitor vulnerable populations, and keep track of the development of chronic diseases, and the transmission of infectious illnesses – such as during pandemics. However, there are many challenges to monitoring patients using wearables, with massive data generation and battery power consumption being significant constraints. Strategies to reduce data generation should be applied taking into account the patient’s clinical status and health risks. Previous studies took advantage of single early-warning scores (EWS) utilized in infirmaries to detect emergencies, reduce transmissions, and be a reference for self-adaptive features embedded in the devices. Our work proposes the use of composite EWS to infer health deterioration risk, minimize data transmissions and power consumption, and reduce excessive alarms through self-adaptive features based on these scores. We also compare our method with previous studies using real patient data. Further, we propose applying self-adaptive features to sampling, processing, and transmission rates. Our method demonstrated enhanced data reduction, 81% fewer readings than the baseline, significant pruning of the number of alarms, and dynamic and automatic inference of patient risk.

Spatial-temporal data analysis helps uncover value from data that moves through space and time. One such data analysis technique is clustering, which groups data based on a distance function to identify outliers or assist in classification tasks. Once spatial-temporal data is clustered with respect to space and time, cluster relationships can be observed, such as clusters entering or leaving another, merging, or splitting. A cluster lifetime describes the relationships that a given cluster had from its start to finish. The set of all cluster lifetimes that are related by the relationships describe a cluster dynamic. In this paper, we report our work in progress on a graph-based analysis approach to cluster lifetime dynamics. We discuss how cluster dynamics can be represented using graphs and the opportunities resulting from this approach, including visualization, graph pattern mining, graph classification, and graph compression. Enabled by graph-processing techniques, the proposed approach facilitates tasks such as the detection of regions of significant increase or decrease in the number of cluster elements (e.g. traffic jams), the calculation of a rise or decay parameter to describe this behavior for classification or comparison tasks, and the identification of a cluster’s lifetime, direction, and distance from or to a given point of interest.

Recent advances in software and artificial intelligence technologies have led to the increasing need to support the collaboration between humans and systems in various application domains. The growing capacity for systems has leveraged the power of building applications that have higher autonomy. However, a proper understanding of allocating tasks to either humans or machines is still lacking, and no suggestions are provided to support this allocation. Current approaches do not consider knowledge about the appropriate level of automation (LOA) in this collaboration and do not support adaptive automation, especially task assignments during the system’s operation. The knowledge about which factors affect the variability in human-system interaction LOA has not been explicitly captured. This paper presents a preliminary study that identifies the factors that influence levels of automation in autonomous systems and present the identified factors as a list. Identifying the factors that influence the level of autonomy of systems advances research in the design of autonomous systems by introducing an adaptive automation approach that can recommend levels of automation to support human-computer interactions. Modern systems must be prepared to identify, capture and process the significant volume and variety of data related to the factors that might influence the variability of systems’ behaviours.

Data science projects involve a machine learning (ML) process based on data, code, and models that change over time. For example, the datasets may increase in size and allow an ML model that requires larger datasets to be applied. However, the dynamic factors that influence model selection are not well understood and explicitly represented. This paper presents ongoing work on an adaptive method for ML model selection in big data science projects. The proposed method involves (i) identifying the factors that affect model selection based on heuristics proposed in the literature; and (ii) modeling the variability of these factors using a feature diagram and constraints that trigger adaptive reconfiguration, that is, changes in model selection due to changes in the variability factors. The applicability of the method is demonstrated through an illustrative use case. The proposed method can lead to an improved understanding of dynamic factors that influence model selection, how these factors explicitly affect the selection, and how the adaptive factors can be represented and automated. This improved understanding can result in a project model selection process that is less implicit and more efficient, more adaptive and explainable, and ultimately constitute a foundation for the creation of novel dynamic software product lines to support this process.

Software development is a collaborative effort in which developers often share knowledge by interacting with artifacts and among themselves. When developers interact with artifacts, what we call a Developer-Artifact interaction, they access or define pieces of information within artifacts. When they interact among, what we call a Developer-Developer interaction, they exchange information using a collaborative platform to clarify an issue, promote an idea, or share a comment. However, the high number of such interactions makes it very difficult to capture and assess the evolution of the developers' knowledge about specific software project artifacts and tasks. On one hand, the knowledge they have decreases over time due to the natural limitations of human cognition that restrict their capabilities to cope with information overload. On the other hand, the more they know about specific project elements, the more they are apt to collaborate. In this paper we describe ongoing work on knowledge-oriented and graph-based models that capture the evolution of developers’ knowledge about software project elements such as artifacts, tasks, similar tasks, and the whole software project, and explore the associated rich project-related connected networks.

There is a lack of approaches and tools to support the development of projects in which humans and machines (e.g., machine learning algorithms) need to collaborate to achieve a specified goal. Specifically, given a set of software development tasks to develop a project collaboratively, how can these tasks be assigned to humans or machines to perform each task most efficiently and effectively? Such understanding is essential to support new methodologies for developing human-in-the-loop approaches in which machine learning automated procedures assist software developers in achieving their tasks. This paper describes our work in progress towards providing an approach to guide the assignment of tasks in developing human-in-the-loop big data (science) software development projects. The paper provides several contributions, including the provision of (i) a human-in-the-loop approach for the development of big data software development projects; (ii) the application of the approach to two case studies; (iii) a discussion of implications and research opportunities.

Internet of Things (IoT) applications have been deployed in several domains, including health care, smart cities, and agriculture. Because of the complex static and dynamic variability of the environment in which these applications are deployed, machine learning-based approaches have been used to support the design of IoT applications. In particular, an emergent approach involves using neural networks to enable IoT devices to learn to adapt their behavior based on the dynamics of the environment. Designing IoT systems is already challenging because of the autonomy and concurrency inherent in distributed physical systems. Moreover, neural networks systems have particular characteristics, such as dynamism, adaptability, and generalization, that make it necessary to adapt the traditional software development lifecycle to satisfy the requirements of these systems. In this paper, we describe our proposed approach to support the engineering of IoT neural network-based systems. Our approach considers a lifecycle supporting the integration of IoT system development tasks with particular ANN tasks, as model requirements and feature engineering. In addition, the paper includes the provision of the application of the approach to a case study and conclusive remarks.

Emergent software applications are increasingly becoming (self-)adaptive and autonomous. Further, Internet of Things (IoT) applications increasingly involve data analytics. The introduction of neural networks in IoT systems has enabled a new generation of applications capable of performing complex sensing and actuation analysis tasks that were not previously possible with other approaches. A key component in the development of these systems is the ability to represent data analytics variability, which captures the ways in which the system can adapt in terms of the data analysis at design and run times. Although variability has been explored in the domain of software product lines (SPLs), data analytics variability in IoT neural network-based systems still seems to be poorly understood and needs to be investigated appropriately. In this paper, we introduce an approach to capture data analytics variability in IoT neural network-based systems (IoTNNSs). The approach represents several types of variability inherent in the development of these analytics systems, including those related to the application context, behavior, quality attributes, IoT devices, and neural networks.

The Internet of Things (IoT) has enabled novel solutions for monitoring patients’ health through wearable sensors in conditions of both non-communicable and infectious diseases. In this paper, we report work in progress involving the development of an IoT-based COVID-19 health monitoring system that can effectively monitor the essential physiological functions of a patient through wireless sensors, thus supporting the early detection of severe cases and the continuous assessment of the patient status. The work provides several main contributions, as it includes: (i) a brief description of the current IoT-based system for remote monitoring of COVID-19 patients; (ii) a description of embedded characteristics of our device, including its contextual functions, early warning score mechanisms and self-adaptive features; and (iii) a description of our preliminary experiment results. Our proposed solution reduced drastically the amount of redundancy in data and still maintain monitoring accuracy. Given the COVID-19 scenarios, in which human resources are extended to the limit and the number of patients in severe conditions is often high, a system that can support IoT-based continuous monitoring are essential to identify changes in clinical status promptly and accurately and can potentially transform the way patients are monitored.

Spatial-temporal data analysis aims at uncovering useful insights and patterns from data that is associated with a location and that changes with time. Spatial-temporal data can comprise massive datasets obtained from multiple sources, including mobile devices, cameras, radar, and other types of sensors. Traditional analysis techniques allow researchers to perform several tasks, including classification, regression, and clustering. Specifically, clustering methods have been widely adopted in domains such as transportation, smart cities, and astronomy. However, current clustering techniques fail to analyze a moving cluster from its start to finish, limiting themselves to investigating static clusters. This study introduces a framework that takes into consideration the entire life of a mobile cluster and describes its lifetime based on dynamic spatial-temporal relationships that the cluster has with other clusters or trajectories. The framework is evaluated using two case studies, which involve taxi trajectories and human mobility.

Data analysis involves the use of a wide variety of systems and libraries to support the exploration and development of models that can uncover valuable patterns and enable individuals and businesses to draw informed insights. However, efforts towards the automation of the ML-based data analysis modeling process faces numerous challenges. In this paper, we describe our ongoing work towards the automation of the data analysis modeling phase based on a variability-aware approach. This approach involves capturing the variabilities through feature models, designing an automated framework to support the analysis, and developing use cases. The work advances the state of the art in the development of methods and tools to support the automation of ML-based data analysis.

Public health monitoring methods have limitations that affect the quality of data. To support traditional data collection efforts, personal smart technologies can be used to collect multimodal, real-time and continuous data. Public health agencies can then study and predict the prevalence of conditions in a population using advanced analytics. Apple Health is one of the most popular sources of health data from personal devices, supporting diverse sensors that collect a wide range of information from heart rate to blood pressure and sleep. This paper introduces a system that uses a mobile health platform to extract Apple Health data to support public health monitoring. Development, security and privacy considerations are discussed, and a pilot study is proposed which collects several objective sensor data from Apple Health as well as self-report perceived stress (both using the platform) to create stress prediction models. Ultimately, the system described can provide public health agencies with novel methods to collect multimodal data from consumer devices as well as implement interventions in real-time to minimize the impact of conditions, such as stress, in a population. The system advances the state-of-the-art in health monitoring by being one of the first works to leverage health data from consumer-level personal devices for public health.

Embodied agents is a term used to denote intelligent agents, which are a component of devices belonging to the Internet of Things (IoT) domain. Each agent is provided with sensors and actuators to interact with the environment, and with a `controller' that usually contains an artificial neural network (ANN). In previous publications, we introduced three software approaches to design, implement and test IoT embodied agents. In this paper, we propose a reference model based on statecharts that offers abstractions tailored to the development of IoT applications. The model represents embodied agents that are controlled by neural networks. Our model includes the ANN training process, represented as a reconfiguration step such as changing agent features or neural net connections. Our contributions include the identification of the main characteristics of IoT embodied agents, a reference model specification based on statecharts, and an illustrative application of the model to support autonomous street lights. The proposal aims to support the design and implementation of IoT applications by providing high-level design abstractions and models, thus enabling the designer to have a uniform approach to conceiving, designing and explaining such applications.

Spatial-temporal data analysis relates to the application of data analysis techniques to data where space and time are both relevant. Usually, the results of these techniques are used to classify or predict a phenomenon, but little attention is given to the explanation of how such phenomenon happened. For example, one may predict that a sporting event will happen at a particular location and date, but little is known about the indications that such event will happen (e.g. a higher number of vehicles on certain streets, parking lots becoming full, large number of vehicles going to supermarkets, or a sudden drop in pedestrian and vehicular traffic movement when the match starts). In this paper, we report on our ongoing work on using spatial-temporal cluster relationships to identify cluster lifecycles. These lifecycles are a series of stages through which a cluster passes during its lifetime, much like a human lifecycle of birth, growth, reproduction, and death. We focus on the identification of cluster lifecycle stages, namely start, expand, shrink, and end, and on their use to predict spatial-temporal phenomena, such as traffic congestion, human events, or animal movement.

Software development is a complex endeavor that depends on a wide variety of contextual factors involving a large amount of distributed information such as technology-related tasks, software operating environments and stakeholder requirements. Most of this context is implicit and captured in the developers' minds (tacit) or distributed through volumes of documentation. Developers have to maintain mental models of this variety of tasks and information as they produce the software. As a result, context can be easily lost or forgotten and developers often use adhoc approaches while finishing the project. We present in this paper the preliminary results of a study that aims at analyzing qualitatively whether supporting software developers with a chatbot during task execution can improve the overall development experience. The chatbot can assist the developers in executing different tasks based on implicit contextual information. We propose an implementation to explore the viability of using textual chatbots to assist developers automatically and proactively with software development project activities that recur. We believe that understanding the interaction of developers with the systems supported by chatbots is key to improving the developer experience and advancing software engineering practices by providing needed timely support for developers.

Innovation depends on the exploitation of market potential with products that are aligned with customer needs. However, building innovative products is becoming gradually more challenging because of increased market volatility, uncertainty, complexity and ambiguity. In our Innovation Lab, inside an e-Procurement Company, we encountered several challenges when implementing an innovation process to develop an initial unstructured data processing Minimum Viable Product (MVP) based on opensource big data tools: (i) raising and prioritizing user demands; (ii) deciding about adequate tools; and (iii) understanding how to promptly set up a viable product. In this paper, we share our open-source innovation experience in bringing novel solutions to an oil company's suppliers. In general, we present and discuss how we have been applying our innovation process to create MVPs, and which technical decision helped us accelerate the MVP development in the presence of a large-scale, unstructured database and open-source big data tools. Overall, we believe the proposed lean-based development process can help practitioners and researchers who want to understand and improve their knowledge about lean products, how to build MVPs, and advance open-source innovation involving big data tools.

Autonomous vehicles (AVs) are vehicles able to sense their environments and guide themselves with little or no human input. AVs have the potential to provide a wide variety of benefits to road traffic, such as improving traffic flow, alleviating traffic congestion and reducing car accidents. However, there are scenarios where AVs require governance systems capable of performing collaborative actions in order to keep traffic flowing smoothly and safely. This paradigm shift will support moving from isolated to collaborative autonomous vehicles. Our goal is to create a collaborative governance-based autonomous vehicle approach using adaptive normative multiagent systems to reduce congestion time. With this collaborative perspective in mind, steps should be taken for AVs to improve their collective mobility while not undermining the `social' goals of vehicle platoons and be able to provide benefits such as reduced congestion, increased travel choice and equity, and reduced pollutant emissions.

Cryptocurrencies, such as Bitcoin, are becoming increasingly popular, having been widely used as an exchange medium in areas such as financial transaction and asset transfer verification. However, there has been a lack of solutions that can support real-time price prediction to cope with high currency volatility, handle massive heterogeneous data volumes, including social media sentiments, while supporting fault tolerance and persistence in real time, and provide real-time adaptation of learning algorithms to cope with new price and sentiment data. In this paper we introduce KryptoOracle, a novel real-time and adaptive cryptocurrency price prediction platform based on Twitter sentiments. The integrative and modular platform is based on (i) a Spark-based architecture which handles the large volume of incoming data in a persistent and fault tolerant way; (ii) an approach that supports sentiment analysis which can respond to large amounts of natural language processing queries in real time; and (iii) a predictive method grounded on online learning in which a model adapts its weights to cope with new prices and sentiments. Besides providing an architectural design, the paper also describes the KryptoOracle platform implementation and experimental evaluation. Overall, the proposed platform can help accelerate decision-making, uncover new opportunities and provide more timely insights based on the available and ever-larger financial data volume and variety.

As a research software platform, iEnvironment has been proposed to support open and big data sharing and reuse for researchers working on surface water issues. Research software refers to software development tools that accelerate discovery and simplify access to digital infrastructures. Although research software platforms are becoming increasingly more innovative and powerful, this increasing complexity hides a greater risk of failure as unplanned and untested program scenarios arise. As systems age and are maintained by different programmers the risk of a change impacting the overall system increases. In contrast, systems that are built with less emphasis on program code and more emphasis on the metadata that describes the application can be more readily changed and maintained by individuals who are less technically skilled but are often more familiar with the application domain. Such systems can also be tested using automatically generated testing regimes.

Spatial-temporal data refers to potentially massive amounts of data gathered across both space and time. Spatial-temporal data analysis helps uncover the value that this type of data holds to domains such as transportation operations, traffic management, service demand, and trip planning. Specifically, cluster analysis groups data into sets known as clusters such that elements inside a cluster are more similar to each other than elements in other clusters. Cluster analysis has been successfully applied in domains such as transportation, ecology, medicine, and astronomy. However, current cluster analysis techniques limit themselves to static cluster analysis, thereby missing the identification of interesting insights and patterns related to the evolution of clusters over time. In this paper, we clarify the concept of dynamic clusters and support new forms of cluster analyses by introducing, describing, and formalizing cluster relationships that represent important events, such as split or merge, that a cluster may go through from its start to its end. These relationships provide a foundation for investigating cluster evolution and providing novel insights for better operational and business decision making.

Software development is a complex activity, which depends on diverse technologies as well as expertise in applying them. Approaches to developing software depend highly on these different characteristics and also the context in which the developers operate. This context involves a massive amount of knowledge, which can often be reused but is easily lost if not captured explicitly and properly organized. Although extensively researched, this context for software development is still not explicit, nor provided as a framework that broadly supports the situation surrounding the software developers and the tools being used. Thus, the ability to reuse this rich context throughout a project is severely hampered. This paper provides a literature review of documents describing the context for software development in traditional and big data projects and a preliminary context-augmented framework design. The purpose of this review is to make the context discovered by the project's software developers explicit and supported by a framework that aids software development using context information. We believe supporting contextual knowledge through its representation and mining for recommendation and real-time provision can significantly improve traditional and big data software project development.

The software development workflow typically involves developers executing tasks and manipulating artifacts.When developers receive a new task they typically envision a task context with the artifacts they intend to manipulate based on their past experiences.Given software projects may last several months, accumulating a vast amount of tasks, artifacts and developers, envisioning this initial task context may be difficult and error-prone.Developers have to walk-through months of past experiences or examine the experience of other developers, select similar tasks and then define the initial context.This paper introduces a method that helps developers defining the initial task context by combining interaction information over artifacts with text information of tasks.First, the Method uses the Clustering technique to organize project tasks into similar groups by interaction in artifacts.Then, the Method uses the Natural Language Processing technique to associate a new task with groups of similar tasks by interaction.The evaluation shows that the clustering of similar tasks by interaction produces similar tasks assigned with artifacts that will be edited by new tasks.The association of new tasks with similar groups by interaction indicates correlation between textual similarity and interaction similarity.

Software development depends on diverse technologies and methods and, as a result, software development teams often need to handle issues in which team members are not experts.To address this lack of expertise, developers typically rely on information obtained from web-based Q&A sites such as Stack Overflow, a popular platform to find solutions to specific technology-related problems.However, access to these Q&A websites is currently not explicitly integrated with software development projects.Therefore, software developers often need to search for solutions to similar and recurring issues multiple times.This lack of integration not only hinders the reuse of the knowledge obtained but also compels developers to perform repeated searches for recurring problems.In this paper, we investigate an approach that explicitly associates project tasks with Stack Overflow posts that have already been curated by developers, and use project task similarities to investigate the possibility to suggest curated Stack Overflow posts.Precision and accuracy were 71.60% and 77.78%, respectively.We also found indications that project task elements such as the process activity, influence accuracy and precision if attempting to reuse curated Stack Overflow posts.

Planning and monitoring the execution of software development activities in agile environments are not trivial procedures.One of the main flaws of agile planning is not considering the dependencies that exist between project tasks.Dependencies between tasks found in software development project plans may lead to the emergence of critical paths, where tasks need to be handled in a strict sequence because the completion of some tasks depends on the completion of others.Not managing such critical paths may reduce team performance and delay product delivery.We performed a study to demonstrate that not identifying dependencies may impair team performance and even increase the risk of costly project delays.The study is divided into three parts: (1) an exploratory study performed in the industry; (2) the development of a tool, AgileCritPath, as a way to support development teams in identifying critical project tasks; and (3) an in vivo evaluation of AgileCritPath based on the Technology Acceptance Model (TAM).The results of the exploratory study provided empirical evidence that there is a need to identify and control dependencies between the tasks in the development of software in agile environments.Using the AgileCritPath tool, allowed us to introduce the Critical Path Method concepts in an agile software development organization.Moreover, the in vivo evaluation demonstrated the benefits of managing tasks dependencies.

This paper describes the development of iEnvironment, an open science software platform that supports monitoring and modeling of aspects of surface water. The platform supports science and engineering research, especially in the context of the creation, sharing, analysis and maintenance of big and open data. In this era of big data, iEnvironment facilitates access to open data resources and research collaboration among science and research disciplines supported by computer scientists and software developers.

Several papers have recently contained reports on applying machine learning (ML) to the automation of software engineering (SE) tasks, such as project management, modeling and development. However, there appear to be no approaches comparing how software engineers fare against machine-learning algorithms as applied to specific software development tasks. Such a comparison is essential to gain insight into which tasks are better performed by humans and which by machine learning and how cooperative work or human-in-the-loop processes can be implemented more effectively. In this paper, we present an empirical study that compares how software engineers and machine-learning algorithms perform and reuse tasks. The empirical study involves the synthesis of the control structure of an autonomous streetlight application.

Agent-based Internet of Things (IoT) applications have recently emerged as applications that can involve sensors, wireless devices, machines and software that can exchange data and be accessed remotely. Such applications have been proposed in several domains including health care, smart cities and agriculture. However, despite their increased adoption, deploying these applications in specific settings has been very challenging because of the complex static and dynamic variability of the physical devices such as sensors and actuators, the software application behavior and the environment in which the application is embedded. In this paper, we propose a modeling approach for IoT analytics based on learning embodied agents (i.e. situated agents). The approach involves: (i) a variability model of IoT embodied agents; (ii) feedback evaluative machine learning; and (iii) reconfiguration of a group of agents in accordance with environmental context. The proposed approach advances the state of the art in that it facilitates the development of Agent-based IoT applications by explicitly capturing their complex and dynamic variabilities and supporting their self-configuration based on an context-aware and machine learning-based approach.

Cluster analysis has helped to uncover changes over time in numerous studies on the dynamics of entities such as people and groups of animals in areas such as human mobility, health, transportation, commerce, and ecology. However, there is a lack of methods that focus on aspects related to the cluster lifecycle, including dynamic analyses on how clusters are formed, change, and disappear. Specifically, how objects enter and exit from the clusters, and how clusters are (de-)composed to form new clusters. In this paper, we introduce our work in progress about an approach to trajectory cluster lifecycle analysis based on big data that supports an evolutionary analysis of clusters throughout their lifecycle. The knowledge that can be captured as a result of such novel forms of analysis will advance the state of the art in a wide range of applications that require information about cluster evolution, and thus provide deeper insights on cluster genesis, existence, and disappearance.

Novel forms of data analysis methods have emerged as a significant research direction in the transportation domain. These methods can potentially help to improve our understanding of the dynamic flows of vehicles, people, and goods. Understanding these dynamics has economic and social consequences, which can improve the quality of life locally or worldwide. Aiming at this objective, a significant amount of research has focused on clustering moving objects to address problems in many domains, including the transportation, health and environment. However, previous research has not investigated the lifecycle of a cluster, including cluster genesis, existence, and disappearance. The representation and analysis of cluster lifecycles can create novel avenues for research, result in new insights for analyses, and allow unique forms of prediction. This paper focuses on studying the lifecycle of clusters by investigating the relations that a cluster has with moving elements and other clusters. This paper also proposes a big data framework that manages the identification and processing of a cluster lifecycle. The ongoing research approach will lead to new ways to perform cluster analysis and advance the state of the art by leading to new insights related to cluster lifecycle. These results can have a significant impact on transport industry data science applications in a wide variety of areas, including congestion management, resource optimization, and hotspot management.

The massive amount of current data has led to many different forms of data analysis processes that aim to explore this data to uncover valuable insights. Methodologies to guide the development of big data science projects, including CRISP-DM and SEMMA, have been widely used in industry and academia. The data analysis modeling phase, which involves decisions on the most appropriate models to adopt, is at the core of these projects. However, from a software engineering perspective, the design and automation of activities performed in this phase are challenging. In this paper, we propose an approach to the data analysis modeling process which involves (i) the assessment of the variability inherent in the CRISP-DM data analysis modeling phase and the provision of feature models that represent this variability; (ii) the definition of a framework structural design that captures the identified variability; and (iii) evaluation of the developed framework design in terms of the possibilities for process automation. The proposed approach advances the state of the art by offering a variability-aware design solution that can enhance system flexibility, potentially leading to novel software frameworks which can significantly improve the level of automation in data analysis modeling process.

The rapidly changing environments in which companies operate to support the Internet of Things (IoT) and Autonomous Vehicles is challenging traditional Multi-agent System (MAS) approaches. The requirements of these highly dynamic environments gave rise to Adaptive Normative MAS approaches. At the same time, governance is an essential and challenging feature that still needs to be addressed in adaptive normative MAS. Indeed, governance of individual and societal agent behavior in Adaptive Normative MASs is still a vague concept that has not been properly investigated, modeled and implemented. However, governance is fundamental for solving problems involving MAS coordination, organizations and institutions. In this paper, we present our ongoing research towards understanding and improving governance in Adaptive Normative MASs. We also discuss challenges and future directions that will facilitate the development of domain-specific smart IoT systems with governance features.

This paper describes iEnvironment, a software platform that supports monitoring and modeling of aspects of surface water. This platform supports surface water science and engineering research, especially when managing and analyzing big and open data. The system will facilitate research collaboration among science and engineering disciplines supported by computer scientists and software developers. In this paper, we briefly describe the iEnvironment architecture, highlighting its reusability and extensibility and some applications.

There is a large amount of discussion in the literature about smart cities where the focus of the discourse is on gathering and analyzing real-time data from smart buildings, smartphones or other sensors to support public services such as vehicular traffic flow, utility consumption or to infer human behaviour. There does not appear to be any discussion of `socially' smart cities where the focus is on using citizens as `smart sensors.' Here the citizens' interactions with a city's services are captured in a timely fashion to derive socio-economic indicators about characteristics of the population relevant to sectors such as education, food security, health, housing, community participation, community safety, income levels and government and to use those as a basis for monitoring community well-being or the effectiveness of government, social service and economic policies designed to produce community improvement. This paper provides the motivation for and outline of a model for a city-level socio-economic indicator system to support the socially smart city. The model is designed to support big highly resolute community data securely. However the model is not just about capturing and analyzing the data; the model must include: deciding what data to collect, developing and communicating with community partners who supply the data and creating a governance structure to ensure that relationships with the data suppliers are maintained. The system will accept timely indicator base data from many different city and other sources and operate on that data using various software tools and maps. The data can be combined in various ways to show single indicators and relationships among indicators. In addition, multiple layers of data can be displayed on a map showing various geographic relationships. An initial version of this model and related system to collect city-level social and economic data and display appropriate socio-economic indicators while protecting individual privacy, is being deployed in a mixed urban-rural community in Southwestern Ontario, Canada. The operational site for the model can be found at myPerthHuron.ca.

Context: Big data has become the new buzzword in the information and communication technology industry. Researchers and major corporations are looking into big data applications to extract the maximum value from the data available to them. However, developing and maintaining stable and scalable big data applications is still a distant milestone. Objective: To look at existing research on how software engineering concepts, namely the phases of the software development project life cycle (SDPLC), can help build better big data application projects. Method: A literature survey was performed. A manual search covered papers returned by search engines resulting in approximately 2,000 papers being searched and 170 papers selected for review. Results: The search results helped in identifying data rich application projects that have the potential to utilize big data successfully. The review helped in exploring SDPLC phases in the context of big data applications and performing a gap analysis of the phases that have yet to see detailed research efforts but deserve attention.

Spatial-temporal computing refers to the modeling, management, and analysis of spatial and temporal information. Despite the recent advances in massive data manipulation, software system approaches that support the massive spatial-temporal data integration and analysis still face numerous challenges, including the lack of: (i) a high-level architectural framework for massive data integration and analysis; (ii) explicit integration and analysis abstractions; (iii) representations of integration and analysis resources; (iv) explicit provenance representation; (v) reusability of integration and analysis steps; (vi) reproducibility of studies; and (vii) models to build and customize integration and analysis applications. This paper proposes the design and implementation of a high-level domain-specific architecture for data integration and analysis that supports building applications in the spatial-temporal domain. The proposed approach describes three types of first-class citizens, which include abstractions to represent data sources, analysis models, and integration operations. It also benefits from domain-specific languages (DSLs) for high-level representations. To make provenance explicit, the proposed approach identifies three types of provenance information, namely description, analysis, and execution, which help to address reusability and reproducibility. Finally, this approach also supports a model-driven technique to generate integration and analysis steps.

Norms have become a promising mechanism to ensure that open multi-agent systems (MASs) produce a desirable social outcome.MASs can be defined as societies in which autonomous agents work to achieve both societal and individual goals.Norms regulate the behavior of agents by defining permissions, obligations and prohibitions, as well as encouraging and discouraging the fulfillment of norms through rewards and punishments mechanisms.Once the priority of software agent is the satisfaction of its own desires and goals, each agent must evaluate the effects associated to the fulfillment or violation of one or more norms before choosing which one should be complied.This paper introduces a framework for normative MASs simulation that provides mechanisms for understanding the impact of norms on an agent and the society to which an agent belongs.

Complex collaborative applications often involve multiple sectors and numerous heterogeneous open and private data sources. Although these applications are starting to emerge as a result of scientific and technological advances in integration and analytical data capabilities, there is a lack of

Water resources, which are a common good, are often negatively impacted through changing patterns of land use, land development, resource extraction and climate. Their sustainability through use and management are often shared responsibilities that involve multiple sectors, including federal and pr

Twenty-first century global change is challenging our use and management of all resources. For a community to adapt and yet maintain and even enhance its economy and quality of life, there is a need for collaborative innovation (CI) and related action among concerned members of the community. This paper outlines the concepts of an approach to CI based on dynamic asset-mapping and its support through a web-based technological framework. Based on real-world experience with the framework, it is clear that CI takes many forms and that it is not possible to build a single set of tools to support CI. Rather a framework and a set of meta-tools is needed which can be used to build tailored systems to fit specific situations that arise when collaboration is to occur.

Software patterns are a new design paradigm used to solve problems that arise when developing software within a particular context. Patterns capture the static and dynamic structure and collaboration among the components in a software design. A key promise of the pattern-based approach is that it may greatly simplify the construction of software systems out of building blocks and thus reuse experience and reduce cost. However, it also introduces significant problems in ensuring the integrity and reliability of these composed systems because of their complex software topologies, interactions and transactions. There is a need to capture these features as a contract through a formal model that allows us to analyze pattern-based designs. In this paper, we provide an overview of a formal framework for ensuring the integrity of the compositions in object-oriented designs by providing mathematically rigorous modeling and analysis techniques for object-oriented systems comprising pattern-based designs as the basic building blocks or design components. A case study related to a hypermedia Web-based application has been presented to illustrate our approach in distributed systems. Copyright © 2005 John Wiley & Sons, Ltd.

Object-oriented frameworks are currently regarded as a promising technology for reusing designs and implementations. However, developers find there is still a steep learning curve when extracting the design rationale and understanding the framework documentation during framework instantiation. Thus, instantiation is a costly process in terms of time, people, and other resources. These problems raise a number of questions including: "How can we raise the level of abstraction in which the framework instantiation is expressed, reasoned about and implemented?" "How can the same high-level design abstractions that were used to develop the framework be used during framework instantiation instead of using source code as is done currently?" "How can we define extended design abstractions that can allow framework instantiation to be explicitly represented and validated?" We present an approach to framework instantiation based on software processes that addresses these issues. Our main goal is to represent the framework design models in an explicit and declarative way, and support changes to this design based on explicit instantiation tasks based on software processes while maintaining system integrity, invariants, and general constraints. In this way, the framework instantiation can be performed in a valid and controlled way.

The consistency problems in UML models and related software processes can be very complex. First, although UML supports a modeling process that should yield inter and intraconsistent models, the meaning of the UML dependencies and their specializations is not precisely defined and, for this reason, many inconsistencies may appear in their models and processes. Precise definitions would form a basis for methods to detect and analyze consistency problems related to UML dependencies and relationships, as well as problems related to software processes described in UML. In addition, we are using the UML object constraint language (OCL) to describe constraints related to the framework instantiation process. OCL is recognized as a limited language in some aspects for expressing well-formedness rules. We have presented a method (T.C. Oliveira et al., 2004) that works in the case of consistent rules and consistent instantiation processes.

Objects and agents are abstractions that exhibit Points of similarity, but the development of multi-agent systems (MASs) poses other challenges to Software Engineering since software agents are inherently more complex entities. In addition, a large MAS needs to satisfy multiple stringent requirements such as reliability, trustability, security, interoperability, scalability, reusability, and maintainability. This workshop brings together researchers and practitioners to discuss the current state of the art and the future research directions in software engineering for large-scale MASs. A particular interest is to understand those issues in the agent technology that make it difficult and/or improve the production of complex distributed systems.

This paper motivates and describes a logic-based approach to specifying and reasoning about interfaces and objects that focuses on separation of concerns issues. The approach is based on the abstract design view (ADV), a software design model for object-oriented systems. The model was originally introduced to characterize, in an informal and practical setting, a clear separation between objects, which we called abstract design objects and their interfaces (ADVs). The objects capture the basic concern, while the interfaces capture special concerns such as user interface, control, timing, and distribution. First, we analyze the ADV design model in order to precisely characterize the interfaces, their associated objects, and the relationship between them. Then, we present one possible approach to formalizing interfaces, objects, and the "views-a" relationship. The central mathematical tools used for this purpose are temporal logic and some tools from the category theory. The formal approach is illustrated by examples that show how the interface and related objects and the views-a relationship can be used in object-oriented specifications. We also show how the theory enables the designer to perform relevant analysis activities while modeling with separation of concerns in mind. The theory can be used to derive dynamic and structural properties of the interface objects and the views-a relationship. In particular, we can use the theory to derive global properties of interfaces that capture special concerns from the local properties of their related objects.

We present a set of problem frames specialized to the area of geographic applications. Problem frames describe the elements of common problems, and help a specifier understand and decompose a complex problem into simpler subproblems for which a solution is often known. We provide an overview of the main issues in geographic applications, and present twelve frames that address many of the problems in the area.

This paper describes an approach to component-based software engineering based on a formal description of design patterns. The architectural design information, captured by design patterns, is made explicit and represented in a declarative way, being packaged into tangible artifacts as building block design components in the development process. These design component descriptions can be instantiated, adapted, assembled, implemented, and maintained. Furthermore, we can use these representations to reason about properties related to the combination of design components. We illustrate the utility of our approach through a case study involving various design patterns.

We describe a black box approach to software system evolution where viewpoints are used to alter a design without modifying the original system. A viewpoint is a model, possibly incomplete, of a domain from a given perspective. We discuss the views-an operator and its semantics, and then apply the concepts to the NACHOS file system as a case study. We believe the approach can be applied to existing object oriented software systems, and serve as a guide for developing systems which are more easily maintained. This case study is an example of an application of a more general transformational approach to software design and implementation

Abstract In this paper we present a logical framework for evolving software systems in which the critical aspects of large-scale software systems maintenance are treated by an approach which considers that configuration management is essentially concerned with the management of system changes. The basis of this framework is a formal description of transitions that might be made in the soft-ware change process, the purpose of the formalism is to allow deductions to be made about the validity of such transitions. A generalised formal description of software system architectures from the viewpoint of evolving software systems is also provided in the paper. Reasoning about functional aspects of these descriptions is achieved through the interaction of a theorem prover for the proposed metalogic formalism and a theorem prover for reasoning about sequential programs. In our approach the alterations of software descriptions can be viewed as programming-in-the-large transformations which are stated by means of the descriptive/prescriptive (deontic) features of actions.