Use of a genetic algorithm in the search for a near-optimal shielding design

An optimization method based on genetic algorithm (GA), which is referred as MACroscopic Near-Optimal Shielding design (MACNOS), is proposed for the search for an optimal radiation shield configuration subject to a given set of constraints. In MACNOS, a GA is used to search for the optimal shielding design and the penalty strategy is employed to deal with the constraints. In order to confirm its capability to search for the optimal shielding design, MACNOS is applied for solving a simple problem with regard to radiation shielding optimization of a hypothetical spaceship reactor. The application shows that, keeping the constraints satisfied, MACNOS is able to seek for the shielding design that minimizes the total weight by changing the thickness and the material of the shield. Therefore, it is expected that MACNOS is potentially useful in the search for the optimal design configuration in the conceptual design phase, where the selection of the shielding material and the estimate of the thickness are necessary. (C) 2009 Elsevier Ltd. All rights reserved.