Andrii Artomov successfully passed his candidacy exam
Congratulation to Andrii Artomov who successfully passed his candidacy exam in August
The main goal of Andrii Artomov’s PhD project is to develop an automated framework for designing complex mechanical systems. The framework should allow solving the following tasks:
- Identify the most promising layout (topology) of the designed system.
- Find optimal values of the system’s parameters for maximizing its performance (typically multiple objectives have to be optimized at once).
- Ensure that the designed system is “robust” i.e., insensitive to internal or external noises such as manufacturing deviations.
While there are algorithms available allowing to address the abovementioned tasks, each task is associated with increased computational costs. Thus, simply combining the existing algorithms into a framework could lead to inadequate computational times. That is why the current project focuses on the development of novel efficient algorithms based on reinforcement learning techniques. The idea is to create a computer program that would act as a “virtual engineer”. Such a model will learn through its own experience gained through the trial-and-error process and once trained it will be able to make decisions in an intelligent way hence reducing the number of computations needed to complete the task.