Search

Highlights • Introduction of agile team formation as an optimization problem. • Proposing expert finding models to retrieve candidates for agile teams. • Introduction of coverage, communication and optimality as quality measures of teams. • Proposing retrieval models to rank generalizing specialist (i.e. T-shaped) members. • Evaluation of proposed models on real datasets gathered from StackOverflow. Abstract Agile team formation is an important requirement of software companies. Since the members of an agile team should be generalizing specialists (i.e. T-shaped experts), we need members who are specialist in a specific topic and have general knowledge in other topics of the team. Selecting such members results in an ideal team which is flexible, high-performing and low-cost. In this paper, we define the problem of agile team formation in which given a set of required skills of an agile team, the ideal output is a set of low-cost candidates who can collectively cover the required skills while they can effectively communicate with each other. We propose two retrieval models to address this problem and then we introduce three evaluation measures for assessment. These measures are coverage, communication and optimality. Our experiments on two test collections extracted from StackOverflow demonstrate the efficiency of our proposed models in comparison with several strong baselines. [ABSTRACT FROM AUTHOR]

• Proposing the defuzzification relations of equality constraints with Nec measure. • Offering a new fuzzy network slack based model on the basis of necessity measure. • The proposed relations combine the Nec equality constraints and network SBM. • Using agile supply chain theory to validate the proposed approach. With respect to the necessity (Nec) measure, suggesting by Dubois and Prade, there is a dearth of any procedure or formula to deal with equality chance constraints. Therefore, this study attempted to introduce (Nec) and possibility (Pos) equality constraints in a real-case network non-oriented slack based model problem via agile supply chain (ASC) theory. The efficiency of different levels of agility of a dairy supply chain (SC), using a new fuzzy network Date Envelopment Analysis slack based model (FNDEA-SBM) on the basis of (Pos)/(Nec) measures, has been measured in this paper. The measurement took place in a multi-stage process. Since the agility features were rather qualitative in nature, the measures were likely to be delineated subjectively using linguistic expressions and fuzzy data. The application of (Pos)/(Nec) measures and different confidence levels made decision makers capable of getting information of all efficiency and sub efficiency scores through optimistic-pessimistic lenses and taking appropriate decisions on the basis of different optimistic-pessimistic points of view and confidence levels. The validity of the suggested approach was demonstrated through a real-life case study related to Iranian top twenty dairy firms. The results indicated the success of the suggested approach and computations were straightforward. [ABSTRACT FROM AUTHOR]

Flexible approaches, such as agile, hybrid, and extreme project management methods in the software project environment, are increasingly being used in nonsoftware environments. Nevertheless, only a few methods and no topological, time-related or resource-related indicators or project databases can address projects of a flexible nature. This study proposes a unified matrix-based model (UMP) to consider single-mode, multimode, individual and multiple projects. Based on the UMP, a CMPD is specified to collect 11 existing project databases. In addition, a flexible structure generator (FSG) is proposed to generate flexible tasks and dependencies for analyzing the effects of flexibility on topology and project demand-related indicators. By correlation graphs, the relations between structure-related and demand-related indicators are analyzed. The comparison of the simulated and real-life databases shows that the interpretation range of the indicators is different; however, considering the flexibility, this problem can be resolved. • Matrix-based models are proposed to unify existing project datasets. • Seven existing single and 4 multiple project databases are combined. • A flexible structure generator is proposed to specify agile and hybrid projects. • New, flexibility-dependent indicators are specified. • The effects of structural flexibility for project topology and demands are examined. [ABSTRACT FROM AUTHOR]

This paper investigated an Earth observation scheduling problem for agile satellites under a time window constraint and resource constraints of limited on-board memory capacity and consecutive working time. We assumed that different observation tasks may have priority levels, and the objective is to maximize the total priority of selected tasks. To address the problem, we first presented a detailed problem description and developed a mathematical programming model. Considering the over-constrained feature of the problem, we developed constructive algorithms to solve the problem, which adopt a priority-based sequential construction procedure to avoid conflicts and generate feasible solutions. The proposed sequential construction procedure contributes to eliminating the need for extra constraint handling techniques, and helps to reduce the complexity of feasibility checking. By analyzing the competitive relationship of various resources, we proved the condition of mutual exclusion of time windows and then developed new priority-based indicators to evaluate the benefits and opportunity costs of different positioning decisions, which is a key component to be used in the proposed constructive algorithms. Through extensive computational experiments on various scenarios including real-world data from China’s satellite platform, the effectiveness of the developed constructive algorithms was verified. [ABSTRACT FROM AUTHOR]

As many software development teams have started to adopt agile methods, a vast amount of valuable experiences have been reported on in both academic and industrial knowledge bases. This information has been used through various approaches to guide and help practitioners finding suitable practices for their software development projects. Nevertheless, not many of these approaches could gather the available experiences to make them systematically reusable and help practitioners understanding agile practices in depth. To the best of our knowledge, only one ontology has been created to solve this problem; some limitations related to its quality and usability make it nevertheless unqualified to serve the intended purpose. The aim of this paper is to build an expert system (i.e. an evidence-based tool) to ease agile practices adoption by efficiently and effectively providing information on them. Firstly, we improve the concepts and relationships in the aforementioned ontology and theoretically validate it using a large data-set of agile practices adoption experiences collected through a Systematic Literature Review (SLR). Secondly, we develop a supporting tool having a friendly Graphical User Interface (GUI) allowing to use the ontology as a concrete agile practice knowledge provider. Finally, we empirically validate the enhanced ontology and evaluate the supporting tool using a survey with agile experts. Our supporting tool can help practitioners to decide what practice to adopt, how to adopt it, how to solve practical issues, etc. The ontology and the tool materialize our contribution to the field of systematic agile practices adoption. • Successful agile practice adoption requires an in depth understanding of its content. • Existing reports on agile practice adoption experiences are essentially unexploitable. • An ontology is used as an efficient solution to recycle practices adoption experience. • A corpus-based approach and a survey are used for validating the ontology. • An expert system supports practitioners querying knowledge on practice adoption. [ABSTRACT FROM AUTHOR]

• Influence Diagram accounts for uncertainties affecting lead time estimation. • Interactive tool to visually communicate probabilistic estimates to stakeholders. • Sensitivity analysis identified team velocity to most drastically impact lead time. • Cumulative distribution function can be an aid to prioritize project backlog queue. This research introduces an influence diagram based approach to estimating task lead times for Agile Kanban project management. Derived from the principles of lean manufacturing, Agile methodologies including Scrum, Scrumban, and Kanban are common in the software industry and are spreading to other fields. Many teams estimate task delivery to better manage stakeholder expectations and improve decision making. However, the current technique involves calculating a team's story point completion velocity, requiring hours of weekly effort while overlooking the addition, removal, and reprioritization of backlog tasks. While these factors may be critical in Scrum, teams practicing Kanban experience these conditions frequently due to the emphasis on continuous integration and reprioritization. Our alternative approach applies an influence diagram, or Bayesian belief network, to model the uncertainties affecting the lead time. To partially automate the estimation process, an influence diagram based expert system is developed and populated with data from a practicing Kanban team and used to generate a cumulative distribution function to facilitate the communication of probabilistic estimates. A sensitivity analysis is conducted to better understand how each factor influences lead times. This system can assist Kanban teams' stakeholder communication and reduce estimation workload through a more holistic model of lead times. [ABSTRACT FROM AUTHOR]

Abstract: Knowledge-based systems (KBSs) have been built and practically applied in various contexts for decades. Yet, they still challenge developers by their complexity: Apart from a sound knowledge base they likewise require comprehensive user interface (UI) and interaction design for supporting the task at hand optimally. However, current KBS-Engineering methodologies still mostly focus on knowledge base development, regarding UI/interaction design a more incidental, low priority task. Thus, the UI often is developed in a rather ad hoc manner, thereby neglecting the potential of experimental design and creating reusable solutions. The integration of (rapid) evolutionary prototyping activities with KBS Engineering offers the chance for incorporating UI/interaction design more prominently into the process while at the same time leveraging the overall development task; specifically, we propose Extensible Prototyping as a tailored prototyping approach and we suggest its integration with the Agile Process Model for knowledge-based systems in this paper. For practical support, we further introduce the customized prototyping and knowledge systems engineering tool ProKEt, and we report on practical experiences with applying both the approach and the tool. [Copyright &y& Elsevier]

• The existing algorithms are not agile enough to adapt to different data structures. • ML approaches are compared to determine suitable ones for the search process. • Neural foresting method is proposed to increases the accuracy of the search results. The increasing amount of pressure to digitalize what we are used to consider a conventional data has created a need to analyze, search and process unique data structures in a timely manner. The progressive world has created a justified need not only for a fast query searches but also the most related and meaningful searches with a minimal guidance. This article explores the potential benefits of using unorthodox solutions, including Artificial Intelligence algorithms in processing big data that are unconventional data structures. One of such big data sources was created as part of digitalization of historic heritage materials, documents and artifacts. The article calls out the benefits and disadvantages of some Artificial Intelligence algorithms and explores ways to use a few of those algorithms for the purposes of the digital heritage. It also offers the solution to maximize the potential of the search engine that could be built for digital heritage or any other unstructured data. [ABSTRACT FROM AUTHOR]