عنوان مقاله
الگوریتم بهینه سازی گروه ذرات مبتنی بر ایده بازپخت شبیه شده
فهرست مطالب
خلاصه
مقدمه
الگوریتم هایSa-PSO
آزمایش شبیه سازی
نتیجه گیری
بخشی از مقاله
ایده اساسی الگوریتم بهینه سازی گروه ذرات بازپخت شبیه سازی شده (Sa-PSO) بصورت زیر نشان داده می شود: در ابتدا، بهترین موضوع فرد و بهترین موضوع جهانی توسط قانون متروپولیس پذیرفته شدند، بهترین مورد هیپو با احتمال مقبول واقع شد، تابع هدف مجازاً در حد خاصی بدتر می شود، قانون پذیرش توسط ضریبT تعیین شد،T دمای بازپخت است.
با نزولT ،ناحیه جستجو می تواند اطراف بهترین نقطه باشد، احتمال پذیرفته نقطه بهترین هیپو نیز اندک خواهد بود، زمانی کهT به حد پایینی می رسد،احتمال پذیرفته نقطه بهترین هیپو صفر است، الگوریتم فقط بهترین راه حل را به عنوان الگوریتمPSO اساسی می پذیرد. رابطه بین دمای بازپخت و وزن اینرسی بوجود آمد، وزن اینرسی با دما تغییر می کند، و سپس دقت جستجو بر اساس وزن اینرسی تغییر کرد، بنابراین سرعت جستجو افزایش یافت.
کلمات کلیدی:
Particle Swarm Optimization Algorithm Based on the Idea of Simulated Annealing DONG Chaojun, QIU Zulian (Automation Control Research Institute, School of Electronic & Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China) Summary Particle swarm optimization (PSO) algorithm is a new population intelligence algorithm and has good performance on optimization. After the standard PSO algorithm and the idea of simulated annealing algorithm had been analyzed, the acceptance of Metropolis rule by probability in the simulated annealing algorithm was introduced in the algorithm of PSO. The simulated annealing-particle swarm optimization was presented. Simulation result shows that the disadvantage of getting in the local best point of standard PSO was overcome effectively and the ability of global optimality are toned up. Keywords particle swarm optimization; Simulated annealing; global optimality; Metropolis rule 1. Introduction Particle swarm optimization (PSO), a kind of evolvement-computation technology based on swarm intelligence, was raised by Kennedy and Eberhart who were aroused by the research results about artificial life in 1995. The basic idea of PSO came from the research on the behavior of bird swarms looking for food. Every particle always follows the two best positions—the best position in the whole swarm and itself in iteration computation, so it converges very fast, but there are several shortcomings in PSO below: (1) getting in a local best point easily; (2) It is difficult to deal with the constraints of the optimization problem. From then the inertia coefficient was discussed [2-6] 2. Sa-PSO Algorithms 2.1 Basic PSO Algorithm [2-3, 5-6] There are m particles in a swarm that is in a space of D dimensions, the ith particle’s position in the space is below: Xi = ( , , , ) i1 i2 iD x x L x , i = 1,2,L,m , which is a latent solution. The ith particle’s flit speed is below: vi = ( , , , ) i1 i2 iD v v L v , i = 1,2,L,m , and until now the best position of the ith particle is below: Pi = ( , , , ) pi1 pi2 L piD , i = 1,2,L,m , the best position in the whole swarm until now is below: Pg = ( , , , ) pg1 pg 2 L pgD , the PSO algorithm is below id id id x = x + v (1) ( ) ( ) id id 1 1 id id 2 2 gd id v = v + c γ p − x + c γ p − x (2) Where, i = 1,2,L,m , d = 1,2,L, D , 1 c and 2 c are, respectively, the study coefficients of cognizing and society, and are both positive constants. The relative value of 1 c and 2 c expresses the relative importance-degree of Pi and Pg with evolvement. 1 γ and 2 γ are both random numbers between 0 and 1, [ , ] max max v v v id ∈ − , max v is decided by the user. Equations (1)-(2) can be changed as id id id x = αx + v (3) ( ) ( ) id id 1 1 id id 2 2 gd id v = wv + c γ p − x + c γ p − x (4) whereα and w are, respectively , constraint factor and inertial factor (w >0). Many scholars regard the equations (3)-(4) as the basic PSO algorithm, It is obvious that the PSO algorithm is very simple, but the algorithm itself includes the principles of sociology, psychology and bionomics, so the particle swarm including some simple particles behaves a complex action, which is the main reason that many peoples are sure that the PSO algorithm has a good foreground.
