39 Journal of Architectural Environment & Structural Engineering Research | Volume 01 | Issue 01 | 2018 Journal of Architectural Environment & Structural Engineering Research http://ojs.bilpublishing.com/index.php/jaeser Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jaeser.v1i1.353 *Corresponding Author: Elder Oroski Universidade Tecnol´ogica Federal do Paran´a (UTFPR), Curitiba, Brazil Email: oroski@utfpr.edu.br AbstrAct Heuristic optimization is an appealing method for solving some engineering problems, in which gradient information may not be available, or yet, when the problem presents many minima points. thus, the goal of this paper is to present a new heuristic algorithm based on the Anthropic Principle, the Anthropic Principle Algorithm (APA). this algorithm is based on the following idea: the universe developed itself in the exact way to allow the existence of all current things, including life. this idea is very similar to the convergence in an optimization process. Arguing about the merit of the Anthropic Principle is not among the goals of this paper. this principle is treated only as an inspiration for heuris- tic optimization algorithms. In the end of the paper, some applications of the APA are presented. classical problems such as rosenbrock function minimization, system identi- fcation examples and minimization of some benchmark functions are also presented. In order to validate the APA's functionality, a comparison between the APA and the classic heuristic algorithms, Genetic Algorithm (GA) and Particle swarm Optimization (PsO) is made. In this comparison, the APA presented better results in the majority of tested cases, proving that it has a great potential for application in optimization problems.. ArtIcLE INFO Article history: received: 26 November 2018 Accepted: 12 December 2018 Published: 31 December 2018 Keywords: Heuristic optimization Anthropic principle System identifcation REviEw Anthropic Principle Algorithm: A New Heuristic Optimization Method Elder Oroski 1 * Beatriz S. Pês 2 Rafael H. Lopez 3 Adolfo Bauchspiess 4 1. Universidade tecnológica Federal do Paraná (UtFPr), curitiba, brazil 2. Instituto Federal do Paraná (IFPr),campo Largo, brazil 3. Universidade Federal de santa catarina (UFsc),Florianópolis, brazil 4. Universidade de brasília (Unb), brasília, brazil 1. introduction I n the last decades, optimization problems have moti- vated great improvements in mathematics and engi- neering. Methods like Newton, steepest descendent and Levenberg-Marquardt have made possible the solu- tion of a series of design optimization problems [1] . How- ever, these methods require strong conditions to have their convergence proved, such as availability of gradients, convexity, and so on [2,3] . It is important to point out that in several industrial applications the designer has to deal with some peculiarities such as non-linearity, non-convex- ity, existence of several local minima, presence of discrete