Improving solution characteristics of particle swarm optimization using digital pheromones

13th AIAA/ISSMO Multidisciplinary Analysis Optimization Conference

2010

Self Cite

Programmable Graphics Processing Units (GPUs) have lately become promising means to perform scientific computations. When appropriately formulated, population based algorithms such as Particle Swarm Optimization (PSO) can leverage the data parallel architecture of GPUs dramatically improving the solution efficiency characteristics. Prior work by the authors demonstrated the feasibility for using GPUs for solving multidimensional optimization problems with digital pheromones in PSO using OpenGL Shading Language (GLSL). However, the programmability of GPUs in recent years fostered the development of a variety of programming languages making it challenging to select a computing language and use it consistently without the pitfall of being obsolete or unstable. This especially applies to design industries that aim at reducing investment and maintenance costs on high performance computing and training their designers to use such equipment. Although different GPU computing languages are available, some hardware specific languages are designed to rake in performance boosts when used with their host GPUs (e.g., Nvidia CUDA). On the other hand, a few are operating system specific (e.g., HLSL). A few are platform agnostic lending themselves to be used on a workstation with any CPU and a GPU (e.g., GLSL, OpenCL). This paper attempts to compare the performance of digital pheromone PSO when implemented on different GPU computing languages. Recommendations will be made on a viable platform for searching multi-dimensional design spaces. In other words, the paper aims to be a useful resource for designers aspiring for using GPUs in their optimization processes. KeywordsVirtual Reality Applications Center, data-parallel architectures, efficiency characteristic, hardware-accelerated, high performance computing, multidimensional optimization, programmable graphics processing unit, computer graphics, hardware, parallel architectures, particle swarm optimization (PSO) Prior work by the authors demonstrated the feasibility for using GPUs for solving multidimensional optimization problems with digital pheromones in PSO using OpenGL Shading Language (GLSL). However, the programmability of GPUs in recent years fostered the development of a variety of programming languages making it challenging to select a computing language and use it consistently without the pitfall of being obsolete or unstable. This especially applies to design industries that aim at reducing investment and maintenance costs on high performance computing and training their designers to use such equipment. Although different GPU computing languages are available, some hardware specific languages are designed to rake in performance boosts when used with their host GPUs (e.g., Nvidia CUDA). On the other hand, a few are operating system specific (e.g., HLSL). A few are platform agnostic lending themselves to be used on a workstation with any CPU and a GPU (e.g., GLSL, OpenCL). This paper attempts to compare the performance of digital pheromone PSO when implemente...

Section: Overview Of Digital Pheromones In Psomentioning

confidence: 99%

Performance of Hardware Accelerated Particle Swarm Optimization with Digital Pheromones on Dissimilar Computing Platforms

13th AIAA/ISSMO Multidisciplinary Analysis Optimization Conference

2010

Self Cite

“…Therefore, the attraction factor is computed as shown in Equation (5). Equation (6) computes the distance between the pheromone and each particle in the swarm. Figure 3 shows an example scenario of a particle being attracted to a target pheromone.…”

Section: Attraction To a Target Digital Pheromonementioning

confidence: 99%

“…Figure 3 shows an example scenario of a particle being attracted to a target pheromone. (6) In the figure, the particle will be more attracted to a pheromone with a higher P' value, as opposed to pheromones that are closer but with a lower P' value.…”

Section: Attraction To a Target Digital Pheromonementioning

confidence: 99%

See 1 more Smart Citation

Implementation of Digital Pheromones in Particle Swarm Optimization for Constrained Optimization Problems

49th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference &Amp;lt;br> 16th AIAA/ASME/AHS Ada

2008

Self Cite

This paper presents a model for digital pheromone implementation of Particle Swarm Optimization (PSO) to solve constrained optimization problems. Digital pheromones are models simulating real pheromones produced by insects for communication to indicate a source of food or a nesting location. When integrated within PSO, this principle of communication and organization between swarm members offer substantial improvement in search accuracy, efficiency and reliability. Multiple pheromones are released in the design space, and the strength of a pheromone in a region of the design space is determined through empirical proximity analysis, and. The swarm then reacts accordingly based on the probability that this region may contain an optimum. The addition of a pheromone component to the velocity vector equation demonstrated substantial success in solving unconstrained problems. The research presented in this paper explores the suitability of the developed method to solve constrained optimization problems. A sequential unconstrained minimization technique -Augmented Lagrange Multiplier (ALM) method has been implemented to address constrained optimization problems. ALM has been chosen because of its relative insensitivity to whether the initial design points for a pseudo objective function are feasible or infeasible. The development of the method and results from solving several constrained test problems are presented.

“…This procedure is continued until a prescribed convergence criterion is satisfied. A detailed account of this procedure is fully explained in the previous work by authors 36 , and is not described in this paper to maintain conciseness. Recently, technologies such as hyper threading and multi-core processing 37 have been the main drivers increasing CPU performance as opposed to the addition of more transistors onto a CPU chip.…”

Section: Overview Of Digital Pheromones In Psomentioning

confidence: 99%

Implementation of Digital Pheromones in PSO Accelerated by Commodity Graphics Hardware

12th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference

2008

Self Cite

In this paper, a model for Graphics Processing Unit (GPU) implementation of Particle Swarm Optimization (PSO) using digital pheromones to coordinate swarms within ndimensional design spaces is presented. Previous work by the authors demonstrated the capability of digital pheromones within PSO for searching ndimensional design spaces with improved accuracy, efficiency and reliability in both serial and parallel computing environments using traditional CPUs. Modern GPUs have proven to outperform the number of floating point operations when compared to CPUs through inherent data parallel architecture and higher bandwidth capabilities. The advent of programmable graphics hardware in the recent times further provided a suitable platform for scientific computing particularly in the field of design optimization. However, the data parallel architecture of GPUs requires a specialized formulation for leveraging its computational capabilities. When the objective function computations are appropriately formulated for GPUs, it is theorized that the solution efficiency (speed) can be significantly increased while maintaining solution accuracy. The development of this method together with a number of multi-modal unconstrained test problems are tested and presented in this paper. In this paper, a model for Graphics Processing Unit (GPU) implementation of Particle Swarm Optimization (PSO) using digital pheromones to coordinate swarms within ndimensional design spaces is presented. Previous work by the authors demonstrated the capability of digital pheromones within PSO for searching n-dimensional design spaces with improved accuracy, efficiency and reliability in both serial and parallel computing environments using traditional CPUs. Modern GPUs have proven to outperform the number of floating point operations when compared to CPUs through inherent data parallel architecture and higher bandwidth capabilities. The advent of programmable graphics hardware in the recent times further provided a suitable platform for scientific computing particularly in the field of design optimization. However, the data parallel architecture of GPUs requires a specialized formulation for leveraging its computational capabilities. When the objective function computations are appropriately formulated for GPUs, it is theorized that the solution efficiency (speed) can be significantly increased while maintaining solution accuracy. The development of this method together with a number of multi-modal unconstrained test problems are tested and presented in this paper.