Research‎ > ‎

Personalized Web-Tasking with Situation-Aware Smart Applications (IBM CAS Project)


From the perspective of the engineering of software systems for the personalization of web-tasking, an important requirement is to provide effective mechanisms to automate repetitive and ordinary tasks while hiding the complexity of these mechanisms from the user. Hence, personalized web-tasking with situation-aware smart applications concerns the design and implementation of situation-aware self-adaptive software applications that understand the user situation and their execution environment, to deliver functionality during runtime that assist people in performing tasks with a minimum effort while maximizing user satisfaction.
Research plan
This project comprises the following milestones:

  1. Software-based approaches to task simplification. To understand personalized web-tasking in the context of the smart internet, we must first understand the state-of-the-art of software-based task simplification. During this milestone we propose a set of research challenges towards task simplifications. That is, the automatic decomposition of a web-task into an ordered and logical sequence of web-subtasks. Read Article
  2. Characterization of models requirements for personalized web-tasking. We proposed a conceptual representation of personalized web-tasking, characterize the corresponding modelling requiremts, and studied different modelling approaches in compliance with these requirements. As a result, we mapped the modelling requirements and the modelling approaches to understand the elements that can be applied into personalized web-tasking as well as the gaps. Read our survey soon
  1. Modeling approach proposal for user tasks. Goal-oriented modelling and the General Context Ontology are the main influences in our proposal of runtime models for personalized web-tasking. We propose a personalized web-tasking ontology using OWL2/RDF, and transform such ontology as an RDF graph that can be manipulated at runtime. Such rutime adaptation meets the requierment of context-awareness, and self-adaptation that we envision in the situation-aware software systems to support personalized web-tasking. 
  2. Prototype to demonstrate selected personalized web-tasking featuresAs an illustration of the personaized web-tasking domain, we implement an scenario based on an online grocery shopping task. This scenario implies activities that must be performed periodically and can involve social interactions, as well as interaction of a variety of software applications and devices. 
  3. Social Context modelling and reasoning support. Social context (i.e., the personal context of other users in her social network) is also part of her personal context. Data analytics is relevant to understand this type of information and act upon it to for improving the web-tasking by suggesting better recommendations to execute web-tasks and/or dynamically compose them. We present a characterization of social context and how it might affect personalized web-tasking. 
  1. Models at Runtime (MART) for PWT. MART are representations of the system and its environment, accessible to the system and capable to evolve t execution time. The goal of this year focuses on the development of the supporting infrastructures of the MART for PWT.
  2. Prototype to demostrate the execution of MART. This year we improve our prototype by including dynamic behaviour on the supporting MART of the system.



* GI Dagstuhl 2014
* SEAMS 2014
* CASCON 2013
* CSER 2013