Optimiser Problem Statement
Optimiser Problem Statement
Optimiser Problem Statement
2024
Problem Statement
Motivation
As an engineer, we always try to model the manufacturing process in such a way
that will optimize the output. We often have limited knowledge about the
process know-how, the exact functional form of the response, and the design
variables. Therefore, we perform experiments or simulate the process to gain
insights. However, experiments are sometimes tedious and expensive. One way
to alleviate this burden is to build an approximate model known as the surrogate
model, e.g., Response Surface Method (RSM), Meta Models, and Emulators. A
model is constructed based on modeling the simulator's response to a limited
number of reasonably chosen data points. This approach is also known as
Black-Box modeling.
Aim
The objective of this competition is to optimize one response variable while
simultaneously satisfying specifications on other response variables. This
competition expects the participants to leverage various mathematical and
statistical models to analyze the limited number of experimental or simulated
data points and achieve the desired optimum output.
Teams Specifications
1. A maximum of 2 participants are allowed per team.
2. Students of different institutions can also form a team.
3. At least one team member must be from a Chemical engineering
background.
2: Register for the Optimiser competition by filling the form in the link. You will
AZeotropy’24
be allotted a unique team ID after the registration. Team ID will be sent via
email. https://www.azeotropy.org/optimiser
5. No participant can register in more than one team. In such an instance, both
teams will be subjected to disqualification.
Structure of Competition
Stage 1: Solving the Optimization Problem
We will provide the simulator (Optimiser app) for computing the value of T. You
can give the values of C & R as inputs and get the corresponding value of T as the
output. The link to the simulator and instructions to run the simulator will be
sent to the participants through mail around a week after the competition
AZeotropy’24
launch. It is mandatory for each team to submit a report consisting of their
approach and a code file. You can use Softwares like MATLAB, Python, Excel, R.
Include all the data points used to obtain the desired response in a tabular form
in the report. Your report should be neat, clean & include necessary plots, tables,
calculations, and mathematical equations. You need to number all tables and
figures, state clearly what each figure aims to convey, and add appropriate
supporting textual explanations of the results. Based on submissions, teams will
be shortlisted for the event day competition.
A mail regarding the team’s selection will be sent. Prepare a final PDF file of your
report and a code file & rename them as follows "Optimiser_Team ID". Upload
the report and code files in this submission link:
https://www.azeotropy.org/optimiser
● Only one member from each team should submit the solution
● Multiple submissions are not allowed, and practicing this would lead to
the cancellation of the submission The report should contain the following
details:
Name, contact information of all the team members, and Team ID
The deadline for abstract submission is February 17th, 2024, at
11.59 p.m.
Stage 2: Viva Round
There will be a viva round for the selected teams from round 1.
Detailed information for the same will be conveyed later.
Evaluation Criteria
Stage 1: This round has a weightage of 80%
A. Choosing the right set of data points: [25 marks]
We expect the participants to choose the data points smartly. Think that
every run of the simulator is equivalent to calculating the value of T without
knowing its functional form. Participants’ goal should be to try to minimize
the number of runs.
B. Analysis of the dataset: [40 marks]
When we have limited knowledge about the process, we rely on the available
mathematical and statistical tools. Performing appropriate analysis on the
right datasets can give us a good approximating model for predicting the
output and lead us to the optimum response.
C. Results [15 marks]
For getting the correct set of the total cost function, and corresponding cost
& risk variables values: 5 marks
Within 5% interval of the total cost function: 5 marks
Within 10% interval of the total cost function: 3 mark
Report: 5 marks
D. Bonus
One can run the simulator a maximum of 20 times. But the lesser you run the
simulator, the more you can save your time and resources. To emphasize this
point, participants will be given a bonus as (20-X)*4, where X is the number of
times a particular team has run the simulator. The upper cap for bonus marks
AZeotropy’24
is 20 (You can get a maximum of 20 marks as a bonus)