Algoritmo genético para otimização de parâmetros de acelerador circular de prótons com cavidade cilíndrica ressonante

Abstract

Circular particle accelerators are devices with wide application in medical areas, like in the radioisotopes production or for therapeutic purposes. However, over the last few years, more and more specifics of the projects of these equipments have been observed due to their application, with the purpose of reducing dimensions and costs, but increasing their performance. Although, such equipment follows requirements already present in previous projects, making them not optimized for the proposed practice. Thus the present work deals with evolutionary computation, specifically genetic algorithms (GA), as a generator of parameters optimized for the design of a circular particle accelerator (AC). The AC model with resonant cylindrical cavity was used as the acceleration structure. A physical-mathematical analytical modeling was proposed and implemented computationally, known as the ACEL code. The ACEL code generated data for the GA, both implemented in Matlab® software, with a minimized fitness function. Restrictions were imposed on the fitness function to guarantee all conditions to the problem, obtaining only valid domain solutions. The results show the feasibility of using a resonant cylindrical cavity as an acceleration structure, as well as the complete model of the circular particle accelerator proposed in this work. In addition, the developed AG and the proposed parameters optimized for specific AC projects are highlighted. Basic optimized physical parameters of the following accelerators were presented: AC generator of positron emitting radioisotopes, as well as radioisotopes that suffer decay by electronic capture (ACPR - 35); AC dedicated to the production of carbon-11 radioisotope (ACC - 11); As well as ACs for proton beam therapy at energies of 65 MeV and 250 MeV (ACPT - 65 and ACPT - 250).