This repository contains the codebase representing each experiment conducted in the research project titled "Optimization Algorithm for Agents in Random Conditions." The project received the 2nd award in the field of computer science at YSC 2022 (Young Scientist Competition 2022), organized by NSTDA (National Science and Technology Development Agency).
- Kampanat Yingseree (Team Leader)
- Sirapat Panmoon
- Assistant Professor Dr. Jakarin Chawachat
The mathematical optimization problem is one of the problems appearing in many fields. In this research, we developed an optimization algorithm for problems with random conditions using the concept of natural selection. We designed a simulation model of market environment to test the efficiency of the algorithm. In our experiments, we found that on random and non-random conditions, our algorithm takes on average 11.27% of the time that Brute-Force takes. Moreover, we tested with Differential Evolution and Dual Annealing and the result came out that our algorithm can perform fastest and can find global maxima even for multimodal function. So, we can conclude that the algorithm can work effectively. We use the expected value as the fitness of the optimization objective, and the result of the optimization has led to every scenario of the environment having a high profit. Our algorithm reduces the flaws that may occur in the optimization. And for the simulation model of the market environment can be implemented to support business decisions, especially for small businesses that have less capital and decision path.
Keywords: Mathematical Optimization, Algorithm, Simulation Model of Market Environment