[์ธ๊ณต์ง€๋Šฅ] Genetic Algorithm Examples
ยท
Study/์ธ๊ณต์ง€๋Šฅ
Example1. A simple function optimization, f(x) = x2 on the integer interval [0,31] encoding scheme → binary unsigned integer of length 5 fitness function → ๋ชฉํ‘œํ•จ์ˆ˜ ์ด์šฉ selection scheme → roulette wehel selection ์‚ฌ์šฉ crossover, mutation scheme → one point parameter setting M(population size) = 4 → coin์„ 20๋ฒˆ ๋˜์ ธ์„œ initializationํ•œ๋‹ค. Pc = 1 Pm = 0.001 Evolving process 01101 → 13 ( fitness๋Š” 13*13 = 169) % o..
[์ธ๊ณต์ง€๋Šฅ] Genetic Algorithm ์š”์•ฝ
ยท
Study/์ธ๊ณต์ง€๋Šฅ
์กฐํ•ฉ ์ตœ์ ํ™” (Combinatorial Optimization, CO) VLSI ๋””์ž์ธ์—์„œ์˜ ๋ฐฐ์น˜ ๋ฐ ๋ผ์šฐํŒ… ๋ฌธ์ œ: VLSI ์…€๋“ค์˜ ์ง‘ํ•ฉ์ด ์ฃผ์–ด์ง€๋ฉฐ, ์ด๋“ค์€ ๊ฒฝ๊ณ„์— ํฌํŠธ๋ฅผ ๊ฐ€์ง€๊ณ  ์žˆ๋‹ค. ๋˜ํ•œ ํฌํŠธ๋ฅผ ์—ฐ๊ฒฐํ•ด์•ผ ํ•˜๋Š” ๋„คํŠธ์›Œํฌ์˜ ๋ชจ์Œ๋„ ์ฃผ์–ด์ง„๋‹ค. ์ „์ฒด ๋ฐฐ์น˜ ๋ฐ ๋ฐฐ์„  ๊ฑฐ๋ฆฌ๋ฅผ ์ตœ์†Œํ™”ํ•˜๊ณ  ๊ฐ ์„ ์ด ํŠน์ •ํ•œ ์ƒ์ˆ˜ ์ดํ•˜์ธ ๋ฐฉ๋ฒ•์„ ์ฐพ์•„์•ผ ํ•œ๋‹ค. ์™ธํŒ์› ๋ฌธ์ œ (The Traveling Salesman Problem): ์™ธํŒ์›์€ ์ตœ์†Œ ๋น„์šฉ์œผ๋กœ ๊ฐ ๋„์‹œ๋ฅผ ํ•œ ๋ฒˆ๋งŒ ๋ฐฉ๋ฌธํ•˜๋Š” ๊ฒฝ๋กœ๋ฅผ ์ฐพ๊ณ ์ž ํ•ฉ๋‹ˆ๋‹ค. ์‹คํ–‰ ๊ฐ€๋Šฅํ•œ ํ•ด๋Š” 1๋ถ€ํ„ฐ n๊นŒ์ง€์˜ ์ˆซ์ž์˜ ์ˆœ์—ด๋กœ ๋‚˜ํƒ€๋‚ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋”ฐ๋ผ์„œ ํ•ด ๊ณต๊ฐ„์˜ ํฌ๊ธฐ๋Š” (n-1)!์ด๋‹ค Local Search ๋ถ€๋ถ„ ํƒ์ƒ‰ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ์ด์ „ ๊ฐ’์— ๊ฐ€๊นŒ์šด ๊ฐ’์œผ๋กœ ํ• ๋‹น ๊ณต๊ฐ„ ๋‚ด์—์„œ ๊ฐ’์„ ์ˆ˜์ • ๋ชจ๋“  ์ œ์•ฝ ์กฐ๊ฑด์ด ๋งŒ์กฑ๋  ๋•Œ..