Cellular Autometa Matlab algorithm

Cellular automata and the Simulation of Bioremediation in contaminated soils – Proposal Nouf Abdulghni 02/05/2018 Abstract

Today, contaminated soil is one of the major environmental problems in industrial countries. Among the different techniques that can be applied in order to remove or disinfect soil, Bioremediation (the natural process in which microorganisms clean up contaminated groundwater and soils by degrading different kinds of contaminants, mostly fuel derivatives.) is one of the most economic and less disruptive techniques. In most cases, the degrading activity of indigenous bacteria is stimulated to accelerate natural phenomena by providing appropriate nutrients (e.g., oxygen, phosphorus, and nitrogen). In order to fully understand Bioremediation techniques, several large scale laboratory tests need to be done. Therefore, by creating a macroscopic cellular automata (CA) model, it can help us reach results without going through major in-situ experiments. The choice of macroscopic automata is motivated by the aim to simulate large-scale models, which will describe the bioremediation of contaminated soils. The agent based model is hierarchical and will be structured in three layers (fluid dynamical layer, solute description layer and a biological layer) each layer depending on the others. The layered structure of the model is convenient as it allows one to optimize subsets of parameters in different phases, for example, the parameters of the fluid dynamical layer can be adapted using the results of the contamination experiments only. In this way a large search space can be broken into more manageable portions. I would like to look deeper into this CA model by looking specifically into phenol contaminated soils, monitor dispersion of different kinds of bacteria to optimize their degrading and resistivity schemes, using genetic algorithms. Also improve the model by reducing the number of parameters; to help enhance the results described by the model and ultimately facilitate and improve the effectiveness of in situ bioremediation. Background

My research was heavily based on the scientific article “Applying Cellular Automata to Complex Environmental Problems: The simulation of bioremediation of contaminated soils.” discussed by Gregori, S.Di. The article looks into different attempts to apply CA modeling to any reproducing systems, but most specifically bioremediation. Most attempts of applying CA modeling have dealt with “ microscopic approaches” where the state variables are limited and considered as discrete values. Gregori, S. Di emphasizes a CA model that is similar to the lattice Boltzman model however microscopic. The approach they adopted was to consider each cell as a portion of space that includes serval pores, each cell is assigned a complex state space which may refer to different variables (water content, density of bacteria, chemical concentration. etc.). Classical genetic algorithm operations that are used in the model only acted on the selected zones.

The Model uses the fitness function 𝑓 = #$%

&

# & in the selection property as an optimization

technique, assuming O and S are two M-dimensional vectors whose components are the values, at M time steps. Plan Using genetic algorithms (GAs) found in Netlogo. GAs will Start with generating a random initial population. After evaluating the generated initial population, the genetic algorithm will try to produce new generations of individuals with higher fitness “goodness” (getting a closer agreement to empirical experiments). New generations would be produced by applying the classical genetic operators such as cloning, recombination or selection, crossover and mutation.

Cellular automata and the Simulation of Bioremediation in contaminated soils – Proposal Nouf Abdulghni 02/05/2018 The process will then loop to achieve many successive generations. The general algorithm will describe how organisms move through the median, absorb nutrients to degrade contaminates and how they move to neighboring cells. Some measures will be taken to account in order to prevent the risk of premature convergence to a suboptimal solution which may affect genetic search.

As mentioned before, different phenomena interacted can be grouped in three classes: • The lower layer (fluid dynamical layer), describes the different fluid phases in the

water: o Immobile (can’t flow) o Capillary water o Gravitational water

The rule is induced that all the water is regarded as immobile until a certain threshold is found.

• The second layer, which describes the physical or chemical fate of the solutes. • The third layer is the biological one, and it shows the interaction if biomass with

its environment: growth or micros based on nutrients. CA scheme assumptions:

• No chemical reaction, since chemical reactions will happen in each cell only

(straightforward). • No predator- prey interactions. • No pore clogging (very ambitious). • Two phase flow: air and aqueous solution.

The algorithm will work by the following steps:

1. Generate a list of organisms (all elements of configuration space) 2. Asses the fitness of each organism taking 3. Pick probability mass function 4. Create two children by “crossing over some genes”

Cellular automata and the Simulation of Bioremediation in contaminated soils – Proposal Nouf Abdulghni 02/05/2018

5. Use a randomly generated position between 1< L< n then swap the parents only if the offspring (children) is better, using Netlogo (simple GA):

6. Repeat to get a new size population (k). Refrencess Gregori, S.Di. Serra, R. Villani, M. (1999). “Applying Cellular Automata to Complex Environmental Problems: The simulation of bioremediation of contaminated soils.” Journal of Theoretical Computer Science, P.131-156. https://ac.els-cdn.com/S0304397598001546/1-s2.0-S0304397598001546- main.pdf?_tid=939c726e-1211-11e8-8e05- 00000aacb35f&acdnat=1518672913_6e9d567387b969e60cee00b07f0d4f25 Gregori, S.Di. Serra, R. Villani, M. (1997). “A Cellular Automata Model of Soil Bioremediation.”, P. 31- 54. http://www.complex-systems.com/pdf/11-1-2.pdf Stonedahl, F. and Wilensky, U. (2008). NetLogo Simple Genetic Algorithm model. http://ccl.northwestern.edu/netlogo/models/SimpleGeneticAlgorithm. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL.