Almost all physical and biological systems exhibit random behavior. Some examples of these systems are share market, road traffic, Internet traffic, call center, tollbooth in freeways, hospitals, spreading of disease, high performance cluster computers (HPCC), telecommunication networks, and population dynamics.
For instance, the amount of time a patient spends in a hospital before he/she is treated is random as it depends on various factors such as availability of the doctor, availability of the surgery room, and availability of post-operation room to mention a few.
In this group, we use probabilistic models to parameterize and investigate such systems, simulate them to analyze their behavior and optimize the parameter values to increase the efficiency of the systems. Often the modeling techniques involve large system of differential-difference equations. Finding solutions of such systems involves computing eigenvalues and eigenvectors of huge matrices. Such computations are often impossible to carry out on a personal computer.
Once the modeling of a system is done, a discrete-event system simulation of the system is developed to simulate the system using the associated model. If rare events are involved in the system or the system consists too many random components, then the simulation must be carried out for sufficiently long time in order to achieve reliable results.
This thrust is engaged in developing algorithms for High Performance Cluster Computers (HPCC) such as “Callisto” to carry out performance analysis of systems with random behavior.
Current topics of interests of this thrust are the modeling and simulation of HPCC, Wireless Sensor Networks (WSN), Multi-cast Internet Routers, Urban Road Traffic Lights, and Advanced Queuing Networks.