文件列表:

ppso_v103\license.txt (1317, 2016-08-27)
ppso_v103\oracle.m (2628, 2016-08-30)
ppso_v103\ppso.m (50999, 2016-11-02)
ppso_v103\testPPSO\Hohmann\Hohmann.m (960, 2015-02-17)
ppso_v103\testPPSO\Hohmann\HohmannGa_con.m (99, 2015-02-17)
ppso_v103\testPPSO\Hohmann\HohmannGa_opt.m (74, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann.m (6027, 2016-08-29)
ppso_v103\testPPSO\Hohmann\testPsoHohmann.mat (58705, 2016-08-29)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta10.txt (538, 2015-02-18)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta11.txt (538, 2015-02-18)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta2.txt (530, 2016-08-29)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta3.txt (538, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta4.txt (538, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta5.txt (537, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta6.txt (537, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta7.txt (538, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta8.txt (538, 2015-02-17)
ppso_v103\testPPSO\Hohmann\testPsoHohmann_beta9.txt (530, 2015-02-18)
ppso_v103\testPPSO\Lyapunov_periodic_orbits\Lyapunov_PSO50Tol-12_C300.mat (1023, 2016-08-29)
ppso_v103\testPPSO\Lyapunov_periodic_orbits\Lyapunov_PSO50Tol-12_C302.mat (1166, 2016-08-29)
ppso_v103\testPPSO\Lyapunov_periodic_orbits\Lyapunov_PSO50Tol-12_C304.mat (1215, 2016-08-29)
ppso_v103\testPPSO\Lyapunov_periodic_orbits\propagateLyapunov.m (1587, 2014-11-14)
ppso_v103\testPPSO\Lyapunov_periodic_orbits\testLyapunov.m (3701, 2016-08-29)
ppso_v103\testPPSO\testPso.m (17132, 2016-08-28)
ppso_v103\testPPSO\testPso.txt (1635, 2016-08-28)
ppso_v103\testPPSO\testPso_con.m (573, 2016-08-28)
ppso_v103\testPPSO\testPso_optCon.m (296, 2016-08-28)
ppso_v103\testPPSO\testPso_optCon_forPpso.m (622, 2016-08-26)
ppso_v103\testPPSO\testPso_optRast.m (305, 2015-02-13)
ppso_v103\testPPSO\testPso_optRast_forPpso.m (351, 2016-08-26)
ppso_v103\testPPSO\testPso_testCase2_forPpso.m (134, 2016-08-26)
license.txt (1528, 2016-11-02)

ppso Version 1.03 Find minimum of function using global Particle Swarm Optimization algorithm. % Copyright (c) 2016, Alessandro Peloni % All rights reserved. % Distributed under the BSD 2-Clause License. % % Please refer to license.txt for the full license. Contact Info: a.peloni.1@research.gla.ac.uk =========================================================================== The algorithm has been tested on several examples. List and explanation of those examples are as follows. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Rastrigins function (MATLAB R2014b User Manual. Global Optimization Toolbox - Genetic Algorithm) --------------------------------------------------------------------------- Comparison between the performances of ppso, MATLAB's particleswarm and MATLAB's ga. NON-DEFAULT SETTINGS - Number of runs: 100 - TolFun: 1e-8 - StallIterLimit: 50 - Maximum number of iterations/generations: 500 - Population size: 25 (ppso and particleswarm), 500 (ga) - tolFunSuccess: 1e-3 RESULTS Regarding the average computational time over 100 runs, particleswarm is the fastest one, while ga is the slowest. The number of function evaluations has not been evaluated. Success Rate (one simulation is considered successful if the value of the fitness function is at most tolFunSuccess greater than the known optimum): - ppso: 100% - particleswarm: 83% - ga: 82% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Simple function of two variables - x in +/-500 (MATLAB R2014b User Manual. Global Optimization Toolbox - Particle Swarm) --------------------------------------------------------------------------- Comparison between the performances of ppso, MATLAB's particleswarm and MATLAB's ga. NON-DEFAULT SETTINGS - Number of runs: 100 - TolFun: 1e-8 - StallIterLimit: 50 - Maximum number of iterations/generations: 500 - Population size: 30 (ppso and particleswarm), 100 (ga) - tolFunSuccess: 1e-3 RESULTS Regarding the average computational time over 100 runs, particleswarm is the fastest one, while ga is the slowest. The number of function evaluations has not been evaluated. Success Rate (one simulation is considered successful if the value of the fitness function is at most tolFunSuccess greater than the known optimum): - ppso: 81% - particleswarm: 17% - ga: 66% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Congress on Evolutionary Computation (CEC) 2005 (Suganthan, P.N., Hansen, N., Liang, J.J., Deb, K., Chen, Y.P., Auger, A. and Tiwari, S., "Problem Definition and Evaluation Criteria for the CEC 2005 Special Session on Real-Parameter Optimization", Nanyang Technological University, Singapore, KanGAL Report Number 20052005, May 2005.) --------------------------------------------------------------------------- Only functions 1-6,8-14 have been tested, with D=10. Even though the competition focused on the function evaluations needed to reach the known optimal solutions, here the main focus has been given to the Success Rate. The Success Rates of ppso, MATLAB's particleswarm and MATLAB's ga are compared, together with the Success Rate of the two particle swarm optimizers that competed in CEC2005, which are the following: - 519-PSO (Tasgetiren, F.M., Liang, Y.C., Gencyilmaz, G. and Eker, I., "Global Optimization of Continuous Functions Using Particle Swarm Optimization", The 2005 IEEE Congress on Evolutionary Computation, Sep 2005) - 620-PSO (Liang, J.J. and Suganthan, P.N., "Dynamic Multi-Swarm Particle Swarm Optimizer with Local Search", The 2005 IEEE Congress on Evolutionary Computation, Sep 2005) SUCCESS RATE func ppso particleswarm ga 519-PSO 620-PSO - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1 96% 100% 100% 100% 100% 2 100% 100% 96% 100% 100% 3 0% 0% 0% 0% 100% 4 92% 0% 0% 100% 0% 5 100% 100% 0% 100% 80% 6 4% 12% 0% 0% 100% 8 0% 0% 0% 0% 100% 9 0% 0% 8% 0% 0% 10 0% 0% 0% 0% 0% 11 0% 0% 0% 0% 76% 12 16% 8% 0% 0% 0% 13 0% 0% 0% 0% 0% 14 0% 0% 0% 0% 88% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Lyapunov Periodic Orbits (example from Pontani, M. and Conway, B., "Particle Swarm Optimization Applied to Space Trajectories", Journal of Guidance, Control, and Dynamics, Vol. 33, No. 5, 2010, pp. 1429-1441.) --------------------------------------------------------------------------- As in the reference paper, ppso has been used to find Lyapunov orbits with several values of the Jacobi constant C (C=[3:0.02:3.04]). NON-DEFAULT SETTINGS - Number of runs: 20 (for each value of C) - Maximum number of iterations: 500 - Population size: 50 - Numerical integrator: ode113 (RelTol = AbsTol = 1e-12) RESULTS C x0, [DU] T, [TU] J_{opt} - - - - - - - - - - - - - - - - - - - - - - - - - 3.00 0.768714 4.334959 3.1499e-13 3.02 0.827949 4.860788 3.7400e-13 3.04 0.789616 3.685672 2.1034e-13 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Optimal Two-Impulsive Transfer Between Two Circular Orbits - Hohmann (example from Pontani, M. and Conway, B., "Particle Swarm Optimization Applied to Space Trajectories", Journal of Guidance, Control, and Dynamics, Vol. 33, No. 5, 2010, pp. 1429-1441.) --------------------------------------------------------------------------- Comparison between performances (in terms of success rate and function evaluations) of ppso and MATLAB's ga. A simulation is considered successful if the difference between the total Dv found by the optimizer and the known optimal one (given by the Hohmann-transfer formulation) is less than TolFun. NON-DEFAULT SETTINGS - Number of runs: 100 - Maximum number of iterations: 500 - Population size: 30 (ppso), 100 (ga) - TolCon: 1e-12 - TolFun: 1e-10 RESULTS beta Success Rate (ppso) Success rate (ga) Avr f-count (ppso) Avr f-count (ga) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2 100% 0% 8e3 1e5 3 100% 0% 8e3 5e4 4 100% 0% 8e3 1e5 5 100% 3% 8.5e3 1.9e4 6 100% 6% 8e3 1.8e4 7 100% 7% 8.7e3 3.7e4 8 100% 9% 8.6e3 5e4 9 100% 9% 9e3 4.9e4 10 100% 11% 1e4 1.5e4 11 100% 13% 8.9e3 1.6e4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Bi-impulsive Earth-Apophis Transfer (example from Vasile, M., Minisci, E. and Locatelli, M., "On Testing Global Optimization Algorithms for Space Trajectory Design", AIAA/AAS Astrodynamics Specialist Conference, 2008, Hawaii.) --------------------------------------------------------------------------- Comparison between the performance (in terms of success rate) of ppso, MATLAB's ga and the versions of pso described in the reference paper. NON-DEFAULT SETTINGS - Number of runs: 200 - Maximum number of iterations: 1e5 - Population size: 10 - 20 - 40 (ppso), 100 - 200 - 400 (ga) - Maximum function count: 1e4 RESULTS Name Population size Success rate Notes - - - - - - - - - - - - - - - - - - - - - - - - - ppso 10 25.5% ga 100 22% PSO505 10 35.5% From Ref. PSO507 10 39.5% From Ref. PSO509 10 43.5% From Ref. ppso 20 36% ga 200 44% PSO1005 20 34.5% From Ref. PSO1007 20 42.5% From Ref. PSO1009 20 38.5% From Ref. ppso 40 62% ga 400 53% PSO2005 40 41% From Ref. PSO2007 40 41% From Ref. PSO2009 40 42% From Ref. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ===========================================================================

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：