Assignment 2

Published

March 11, 2024

Due: 03/03/2024

Aim

We have learned Particle Swarm and Genetic Algorithm in the lectures. In this assignment, we will learn a new non-gradient algorithm, implement it, test it and report our findings.

You will learn:

  • how to read and assimilate new optimisation algorithm from research papers
  • how to implement a new optimisation algorithm
  • how to assess the performance of an optimisation algorithm

in this assignment.

Steps

  1. Identify the original paper for the algorithm assigned to you. Read the paper and contact the instructor if you face any difficulty.
  2. Implement the algorithm yourself or modify an opensource code (if available) for your purposes
  3. Test the algorithm on at least 2 problems which you already solved in Assignment 1 using Gradient optimisation. Compare the performance of the algorithms.
  4. If the algorithm assigned to you does not find all the minima, search for modifications of the algorithm which can help it find all the minima as opposed to only the global minima.
  5. Implement the modifications and test the algorithm again on a multi-modal function
  6. Prepare presentation
    • Presentation should not have mathematical equations in the form of copy-paste images from research papers. Typeset all the equations.
    • Latex based presentation slides are preferred but not mandatory
    • Presentation should not contain more than 15 slides (since the presentation is only for 20 minutes)
Note

You are allowed to take help from me or other students. However, the presentation should be your own work. You should not copy the presentation from the internet or from other students. Also the code should be your own work. You can take help from the internet but you should understand the code and write it in your own words.

Deliverables

  • 20 minute presentation
    • You should explain the algorithm in detail
    • You should compare the steps of the algorithm with other algorithms and highlight the similarities and differences
    • Present solutions for at least 2 problems which you already solved in Assignment 1 using Gradient optimisation. Present the number of function evaluations and the number of iterations required to reach the global minimum.
    • Any general observations.
  • Source code implementing the algorithm in a language of your choice. The source code should be well documented.
  • Source code of the presentation slides

Evaluation Criteria

You will be evaluated based on:

  1. your understanding of the algorithm
  2. quality of the presentation (clarity of thought and delivery)
  3. quality of your implementation of the algorithm. Here, we will look at things like the speed of the algorithm and the amount of memory it consumes.

After each presentation, there will be a 10 minutes QA session for everyone.

Algorithm Allocation

Student Slot Algorithm
Akash Differential Evolution
Navya Ant Colony Algorithm
Amritha Cuckoo Search
Nitin Bat Algorithm
Ayushman Firefly Algorithm
Lalitha Harmony Search
Ishmael Artificial Bee Colony Algorithm
Kiran Fish School Search
Vineeth Stochastic Tunnelling
Manjiri Bacterial Foraging
Meetkumar Quantum Annealing
Mohanraj Cross Entropy Method
Ramakrishnan Hill Climbing
Rithik Simulated Annealing
Sidhartha Galactic Swarm Optimization
Suyog Krill Herd
Ayesha Monkey Search
Vishal Runner Root algorithm

There are more algorithms that need to be investigated. If anyone is interested in them, you can also choose from one of these. A good source to find more algorithms is J R R Martin’s wiki.