Cellular Automata as models in Social Sciences
Rezki Chemlal


Abstract.   In this project we were interested in the use of 2 dimension cellular automata in social sciences , we were very flexible in the last point as we also inculed models in urban modeling.We also did not restict ourselevs to the formal definition of cellular automata as we considered also interesting models that may be considered close to cellular automata. We realised programs for all models.

Modeling the Spread of Covid-19 with 2-D Cellular Automata
Subrata Paul
Mentor: Kamalika Bhattacharjee

Abstract.   This project focuses on developing a model that can accurately simulate COVID-19's global dissemination patterns. We utilize cellular automaton (CA) to build such a model. In this project, a new variant of CA called Temporary Stochastic Cellular Automata (TSCA) is used, where two rules is being utilized, one of which serves as a default rule and the other of which is a probabilistic rule that is applied with some probability. To consider the mutation of the COVID-19 rule, we employ a set of TSCAs. Each TSCA is considered as (f, g)[τ] and at a time the applied TSCA is chosen from the set. The model evolves using two TSCA rules f and g, where f is represented as the propagation of the virus, g is represented as recovery function and g is applied with probability τ . The model is validated on the basis of a real-time dataset of spreading Coronavirus (SARS-COVID-19) over the world. This proposed model depicts the spreading scenario of the novel Coronavirus which has caused a global pandemic.

Reversibility in alpha-asynchronous Cellular Automata and Cryptography
Sourasish Das, Mainak Shaw, and Aravind P
Mentor: Souvik Roy

Abstract.   Reversible computation has garnered a lot of attention over the years,Though studied extensively in a great variety of synchronous computation models, it is virtually unexplored in an asynchronous framework. While discussing asynchronous frameworks, alpha asynchronous cellular automata is a niche topic in the discussion of reversibility. Alpha Asynchronous cellular automata update their cells according to the value of alpha, since most of the updation process is based on probability, the experiments to be done to decide on the reversibility was done over a range of CA sizes. Further 88 rules were identified which depicted the characteristics of the 255 rules , and a sample of 10 probabilities were taken and experiments were conducted for all initial configurations for a given CA size and given value of alpha.

Cellular Automata based Modelling of Stock Market
Krishnendu Dasgupta
Mentor: Souvik Roy

Abstract.   In this work, it has been tried to showcase the Dynamics of the Stock Market with the help of a 2D Cellular Automata. The Active Traders are characterised by the states +1(Buying Stocks) and -1(Selling Stocks). The Inactive Traders are characterised by 0 State. Some Simulation Rules(For changing the State of the Traders from Active to Inactive and Vice Versa) are applied and the simulation is done. Based on that Simulation, some Graphs are plotted. After that, Simulation rules are Tweaked. Instead of checking the Neighbours and keeping the condition on the basis of at least 1 Neighbour in the Simulation Rules, 'K' and 'l' Neighbours are respectively checked for the Rule(1) and Rule(2) of Simulation where 'K' and 'l' are taken from the Set {1,2,3,4}. Graphs are obtained and along with that some Drastic Changes are also observed in the Dynamics of the Market. We witness the Strictly Increasing Graphs of the Simulation to become sometimes Strictly Decreasing when the 'K' and 'l' values are Increased. In the later part of the analysis, we take The Global Neighbourhood Condition for the Simulation to resemble the Real Life Stock Market. These Neighbours are Fully Random in nature.

Convergence in 3-state Non-uniform Cellular Automata and Pattern Classification
Aryan Pareek
Mentor : Nazma Naskar

Abstract.   1-dimensional three-neighborhood binary irreversible cellular automata (also called elementary reversible cellular automata or ECA) is a well explored field both with respect to theory and applications. The irreversible behavior of such CA, convergence property for point attractors has been studied in detail for null boundary CA, as well as periodic boundary conditions. The feature set of this cellular automata consists of only binary numbers 0, 1. As a result, we may believe that this may lead to some restrictions of such a pattern classifier in its ability, accuracy, capacity and range of classification. Also, when compared to the real world data, a 2-feature CA really falls short as most data has the number of features spread across the discrete number space.

Finding the Inverse Cellular Automata of a given 1-D Reversible CA
Khitish Kumar Gadnayak
Mentor: Kamalika Bhattacharjee

Abstract.   Reversibility is one of the important chacteristic in the domain of cellular automata since it gives the notion of preserving the information during the evolution. This property leads to give wide aspect of applications in many real life scenarios. The main objective of this project work is to analyse the bijectivity property of the reversible CA with the help of state transition diagram (STD). This report also describes the modelilng approach by mapping of bits of transition states with the rule minterm (RMT) sequences to obtain the inverse function for the reversible and semi-reversible CAs.

Hexagonal Cellular Automata Simulation of Intergranular Cracking in Polycrystalline Materials
Tarun Kumaar M K, Deeraj H, and Suvetha M
Mentor: Prince Gideon Kubendran Amos

Abstract.   The model takes an input of the desired number of nuclei, which are then distributed at random throughout the domain. For this system, we are considering 18 different grain orientations associated with their own unique colours. Conventionally, cellular automata are associated with square discretization but here we have implemented hexagonal discretization and attempt to undertake a qualitative comparison between them.

Understanding The Convergence in Alpha-Asynchronous Cellular Automata
Bheemakonda Raja Maheswar and Revanth D. Rampal
Mentor: Souvik Roy

Abstract.   This paper contains the keen observations and research on factors of convergence, converging rules and attractors of Alpha-Asynchronous Cellular automata. Our experimentation will be on elementary cellular automata from alpha values 0.1 to 1.0 for a set of 88 minimal rules. In this process our main focus is to track converging rules and record them. To get this information we developed a program that takes alpha values and rules as input and returns above mentioned information as output. Our second focus will be on recording the attractors.

Characterization of 2D Linear Cellular Automata by using Linear Algebra
Tina Das
Mentor: Kamalika Bhattacharjee

Abstract.   In this project we studied the behavior of 2D linear Automaton generated by linear rule under the null boundary condition over the field ℤ ² transition matrix or rule matrix of this cellular Automata by represent each cell of block into linear form 1, 0 according to dependency and independency of each cell respectively determined by applied rule.

Isomorphism In Cellular Automata
Vicky Vikrant
Mentor: Sukanya Mukherjee

Abstract.   This work focuses on the isomorphism of cellular automata (CAs). Two cellular automata are said to be isomorphic if their configurations evolve in the similar way. As a model, here we use non-uniform elementary cellular automata under null boundary and discover few inherent properties of CAs to decide whether the given cellular automata are isomorphic.

Multi-seed Image Segmentation for the Microscopic Image
Tanisha Ayach, Suchitra Behera, and Sagarika Padhi
Mentor: Nazma Naskar

Abstract.   Image segmentation is one of the most important image processing techniques for biological images. Till now, various Cellular Automata(CA) rules have been used for segmentation. For updation of states of the cell, conditional rules using Von Neumann neighbourhood and Moore neighbourhood were used. In this work, instead of taking one seed point per image, multiple seed points are used for segmentation. A dataset of microscopic images has been taken and various segementation rules have been used for analysis.