PhD Imran Mahmood

– I am from Lahore in Pakistan. Before my docotoral studies I received my master degree in Distributed Systems from KTH. Before that I have been in the IT industry for around 7 years. My last appointment was in Virtual University of Pakistan as Head of Systems Department. I did my Bachelors in Computer Science (Honors) from the University of Management & Technology Lahore in 2000. 

– My topic is "A Verification Framework for Component Based Modeling and Simulation - 'Putting the pieces together' ”. In my Masters studies I did my internship at Swedish Defense Research Agency (FOI). During my Masters research at FOI, I became interested in the Component based modeling & Simulation and especially about the problem of Composability Verification, as it is known as a hard problem due to its increasing complexity in the system modeling. I therefore decided to continue working in this area.

– To understand it in simple words: Different system Models (we call them components) can be composed together to form complex models. This saves time and cost of development and favors the concept of reuse. However when the components are composed they need to be verified so to ensure that they will perform as intended and according to the user's requirements. It is simple to verify that the components are correctly connected however it is much harder to verify that they will interact with each other in a meaningful way. It is even harder to verify that they will behave according to the given requirements and that they will ultimately reach their goals. 
We require a verification framework that provides tools & techniques to evaluate the composed model of a given set of components and verifies that they can be composed together such that the resultant composed model satisfies the given requirement specifications. 

– In this research we studied various levels of Composability and for each level we proposed a verification technique. To validate our technique we conducted a couple of case studies. In one case we considered a scenario of a Manufacturing System, in which two machines share a common Robot as a resource. Our objective of verification here was to check if they robot behaves fairly in providing service to both machines. Here one of our method successfully evaluated the fairness property of the Robot component. These results were published in a conference paper. In a second case we considered a Field Artillery Military Model in which we modeled the process of Indirect Fire Support. Our objective was to evaluate if they Field Artillery Components  can correctly co-ordinate with each other in order to destroy all the assigned enemy targets (within given time). We use an exhaustive analysis technique, which successfully evaluate the Artillery Model according to its requirement specifications (These results were also published in a conference paper). 
In summary, model verification is not a standard science. It varies depending on the nature of the problem. We explored these variations and try to answer important questions by proposing different methods and techniques in a single framework.

– In the current context our framework is applicable for the Verification of Component Based Simulation Models. We however believe that its application can be extended to other engineering & research areas such as:

– In short, our approach not only brings improvement in the practice of system modeling but also provide robustness in the design by assuring the correctness, consistency and completeness.