Fwd: Fw: Gyan Data Seminar Series

[Danny M]

Dear all:

Please find the attached information about the special seminar hosted by GYAN DATA in the research park on Monday 11 July 2016 at 3:30 PM in Hall 1 in research park.

The speaker will be Dr Mani Bhushan, faculty of IIT Bombay, former PhD student of Prof Raghunathan Rengaswamy.

He will be talking about "Linear Machine: A Novel Approach to Point Location Problem" [For more details scroll down]. The target audience for the talk will be the Systems group in our department. Hence I encourage you to attend the talk.

Regards

Danny Raj M
Doctoral Scholar,

Department of Chemical Engineering

IIT-M, India

ph: +91 9500 195 720

---------- Forwarded message ----------
From: Raghunathan Rengasamy <rag...@iitm.ac.in>
Date: Fri, Jul 8, 2016 at 7:44 PM
Subject: Fw: Gyan Data Seminar Series
To: "danny....@gmail.com" <danny....@gmail.com>

Dear Danny:

  Please send this to Senai with a special note to attend. Also, please send it to the SCG group.

Best wishes

Raghu

 

---

From: Suchitra Veeravalli <ad...@gyandata.com>
Sent: 08 July 2016 15:33
To: Emails for everyone
Subject: Gyan Data Seminar Series

 

Dear all:

We would like to invite you to attend a talk by Dr. Mani Bhushan, faculty, Department of Chemical Engineering, IIT Bombay.

Linear Machine: A Novel Approach to Point Location Problem

 

Speaker:    Dr. Mani Bhushan

Professor

Department of Chemical Engineering

                        IIT Bombay

Date: 11/07/2016

Time: 3.30 pm

Duration: 1hr

Venue: Hall 1, IIT Madras Research Park, Tharamani, Kanagam Road, Chennai 600 113

---------------------------------------------------------------------------------------------------------------------

Abstract: Online implementation of various tasks such as model based
optimal control, state estimation, etc. involves repeated solving of
optimization problems with same structure but with different parameter
values. In recent literature, multiparametric programming has emerged
as a popular option for facilitating online optimization of such tasks for
fast dynamical systems. Multiparametric programming divides the parameter
space into the so-called critical regions and pre-computes optimal
solutions offline as a function of the parameters for these critical
regions. The online step thus involves determination of the critical
region in which the current point(parameter) is located. However, for
large number of critical regions, this point location problem can itself
become computationally challenging.
   In this work, we propose a linear machine based novel approach for the
point location problem. The approach, motivated from pattern classification
literature, involves associating linear discriminant functions with the
critical regions. The current point is then assigned to the critical region
with the largest discriminant function value. We also establish a
necessary condition for existence of linear machine and propose a modified
procedure, involving systematic subdivision of the parametric space, when
this condition is not satisfied. Analysis of complexity and storage
requirements, along with computational experiments on a large sized
example, indicate that linear machine can be an efficient tool for solving
the point location problem.
---------------------------------------------------------------------------------------------------------------------

About the speaker:

Dr.Mani Bhushan is currently a Professor in the Department of Chemical Engineering, IIT Bombay. He earned both his Ph. D. in Chemical Engineering (2001) and his B. Tech. (1997) in Chemical Engineering from IIT Bombay.  Mani was a post- doctoral fellow at Purdue University, from 2001 – 2002.

Dr. Mani Bushan was an assistant professor at the University of Alberta, Canada for a couple of years before returning to his Alma mater, IIT Bombay. His research areas include:

1. Sensor Network Design

2. Optimal Alarm Management

3. Process Planning and Scheduling

4. Process Safety Analysis

 

 

Looking forward to seeing you all on Monday.

Suchitra