1. Trang chủ
  2. » Luận Văn - Báo Cáo

Bài toán quy hoạch toàn phương lồi ngặt với nhiễu giới nội

111 12 0

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 111
Dung lượng 698,63 KB

Nội dung

BỘ GIÁO DỤC VÀ ĐÀO TẠO BỘ QUỐC PHÒNG HỌC VIỆN KỸ THUẬT QUÂN SỰ Bài toán quy hoạch toàn phương lồi ngặt với nhiễu giới nội Võ Minh Phổ Chuyên ngành: Toán học Mã số: 62 46 30 01 Người hướng dẫn khoa học: GS TSKH Hoàng Xuân P PGS TS Phan Thành Anh 2011 `.I CAM D - OAN LO Tˆoi xin cam d¯oan nh˜ u.ng kˆe´t qua˙’ d¯u.o c tr`ınh b`ay luˆa.n a´n l`a m´o.i, d¯a˜ d¯u.o c cˆong bˆo´ trˆen c´ac ta.p ch´ı To´an ho.c quˆo´c tˆe´ C´ac kˆe´t qua˙’ viˆe´t u v`a PGS TS Phan Th`anh An d¯a˜ chung v´o.i GS TSKH Ho`ang Xuˆan Ph´ d¯u.o c su d¯`oˆng y ´ cu˙’a c´ac d¯`ˆong t´ac gia˙’ d¯u.a v`ao luˆa.n ´an C´ac kˆe´t qua˙’ nˆeu luˆa.n a´n l`a trung thu c v`a chu.a t` u.ng d¯u.o c cˆong bˆo´ bˆa´t k` y cˆong tr`ınh n`ao kh´ac tru.o´.c d¯o´ Nghiˆen c´ u.u sinh ˙’ M O.N `.I CA LO Luˆa.n ´an d¯u.o c ho`an th`anh du.o´.i su hu.o´.ng dˆa˜n, chı˙’ ba˙’o cu˙’a GS TSKH Ho`ang Xuˆan Ph´ u v`a PGS TS Phan Thanh An T´ac gia˙’ chˆan th`anh ca˙’m `ay d¯˜a d`anh cho T´ac gia˙’ b`ay to˙’ l`ong o.n su gi´ up d¯o˜ mo.i mˇa.t m`a c´ac Thˆ `ay d¯a˜ u, Thˆ biˆe´t o.n sˆau sˇa´c v`a chˆan th`anh t´o.i GS TSKH Ho`ang Xuˆan Ph´ `eu kiˆe.n d¯ˆe˙’ t´ac quan tˆam, hu.o´.ng dˆa˜n tˆa.n t`ınh, nghiˆem khˇa´c v`a ta.o mo.i d¯iˆ gia˙’ c´o thˆe˙’ ho`an th`anh nh˜ u.ng mu.c tiˆeu d¯ˇa.t cho luˆa.n a´n T´ac gia˙’ xin - oˆng Yˆen, PGS TS Ta Duy b`ay to˙’ l`ong biˆe´t o.n d¯ˆe´n GS TSKH Nguyˆ˜en D Phu.o ng, PGS TS Nguyˆ˜en Nˇang Tˆam v`a c´ac d¯`ˆong nghiˆe.p thuˆo.c Ph`ong Gia˙’i t´ıch sˆo´ v`a T´ınh to´an Khoa ho.c Viˆe.n To´an ho.c v`ı d¯a˜ c´o nh˜ u.ng y ´ kiˆe´n qu´ y b´au cho t´ac gia˙’ qu´a tr`ınh nghiˆen c´ u.u T´ac gia˙’ xin d¯u.o c b`ay to˙’ l`ong ca˙’m o.n d¯ˆe´n Ban chu˙’ nhiˆe.m Khoa Cˆong Nghˆe thˆong tin, Ph`ong Sau d¯a.i ho.c v`a Ban Gi´am d¯oˆ´c Ho.c viˆe.n K˜ y thuˆa.t `eu kiˆe.n thuˆa.n lo i d¯ˆe˙’ t´ac gia˙’ c´o nhiˆ `eu th`o.i gian thu c Quˆan su d¯a˜ ta.o mo.i d¯iˆ hiˆe.n luˆa.n ´an - a`o Thanh T˜ınh, T´ac gia˙’ c˜ ung b`ay to˙’ l`ong biˆe´t o.n d¯ˆe´n PGS TS D -u PGS TS Nguyˆ˜en D ´.c Hiˆe´u, PGS TS Nguyˆ˜en Thiˆe.n Luˆa.n, PGS TS `ong, TS Nguyˆ˜en H˜ Tˆo Vˇan Ban, TS Nguyˆ˜en Nam Hˆ u.u Mˆo.ng, TS V˜ u Thanh H`a, TS Nguyˆ˜en Ma.nh H` ung, TS Nguyˆ˜en Tro.ng To`an, TS Ngˆo -u - `ınh So.n, TS Trˆ `an Nguyˆen Ngo.c H˜ u.u Ph´ uc, TS Tˆo´ng Minh D ´.c, TS Lˆe D v`a tˆa´t ca˙’ c´ac d¯`oˆng nghiˆe.p Khoa Cˆong Nghˆe thˆong tin, HVKTQS, d¯a˜ d¯ˆo.ng viˆen, kh´ıch lˆe v`a c´o nh˜ u.ng trao d¯oˆ˙’i h˜ u.u ´ıch suˆo´t th`o.i gian nghiˆen c´ u.u v`a cˆong t´ac T´ac gia˙’ ca˙’m o.n sˆau sˇa´c GS TSKH Pha.m Thˆe´ Long, Gi´am d¯ˆo´c Ho.c `eu kiˆe.n vˆ `e mˇa.t thu˙’ tu.c c˜ Viˆe.n KTQS, ngu.o`.i d¯a˜ ta.o mo.i d¯iˆ ung nhu chuyˆen mˆon d¯ˆe˙’ t´ac gia˙’ c´o thˆe˙’ ho`an th`anh luˆa.n ´an n`ay Cuˆo´i c` ung t´ac gia˙’ gu˙’.i l`o.i c´am o.n t´o.i vo v`a c´ac con, nh˜ u.ng ngu.o`.i d¯a˜ `eu kiˆe.n cho t´ac gia˙’ qu´a tr`ınh l`am d¯oˆ ng viˆen, chˇam s´oc v`a ta.o mo.i d¯iˆ luˆa.n ´an Mu.c lu.c L` o.i cam d ¯oan L` o.i ca˙’m o.n Danh mu.c c´ ac k´ y hiˆ e.u thu.` o.ng d` ung `au Mo˙’ d ¯ˆ `oi `oi, quy hoa.ch to` a h` am lˆ B` to´ an quy hoa.ch lˆ an phu.o.ng v` thˆ o `oi, quy hoa.ch to`an phu.o.ng 1.1 B`ai to´an quy hoa.ch lˆ `oi suy rˆo.ng thˆo 1.2 H`am lˆ 12 `oi ngo`ai 1.3 H`am γ-lˆ 13 `oi ngo`ai 1.4 H`am Γ-lˆ 15 `oi 1.5 H`am γ-lˆ 17 - iˆ D e˙’m infimum to` an cu.c cu˙’a B` to´ an (P˜ ) `oi ngo`ai cu˙’a h`am bi nhiˆ˜eu 2.1 T´ınh γ-lˆ - iˆe˙’m cu c tiˆe˙’u to`an cu.c v`a d¯iˆe˙’m infimum to`an cu.c 2.2 D 2.3 C´ac t´ınh chˆa´t cu˙’a d¯iˆe˙’m infimum to`an cu.c `eu kiˆe.n tˆo´i u.u 2.4 T´ınh chˆa´t tu a v`a d¯iˆ 20 20 27 28 33 ˜ `oi ngo` T´ınh Γ-lˆ cu˙’a h` am bi nhiˆ e u v` a d ¯iˆ e˙’m infimum to` an cu.c cu˙’a B` to´ an (P˜ ) `oi ngo`ai cu˙’a h`am bi nhiˆ˜eu 3.1 T´ınh Γ-lˆ - iˆe˙’m infimum to`an cu.c cu˙’a b`ai to´an nhiˆ˜eu 3.2 D 43 3.3 T´ınh oˆ˙’n d¯.inh cu˙’a tˆa.p c´ac d¯iˆe˙’m infimum to`an cu.c `eu kiˆe.n tˆo´i u.u 3.4 Du.o´.i vi phˆan suy rˆo.ng thˆo v`a d¯iˆ 55 ˜ - iˆ D e˙’m supremum cu˙’a B` to´ an (Q) `oi cu˙’a h`am bi nhiˆ˜eu 4.1 T´ınh γ-lˆ - iˆe˙’m supremum to`an cu.c cu˙’a h`am bi nhiˆ˜eu 4.2 D 4.3 T´ınh chˆa´t cu˙’a tˆa.p c´ac d¯iˆe˙’m supremum to`an cu.c 4.4 T´ınh chˆa´t cu˙’a tˆa.p c´ac d¯iˆe˙’m supremum d¯.ia phu.o.ng 43 52 58 64 64 66 73 86 Kˆ e´t luˆ a.n chung 94 Danh mu.c cˆ ong tr`ınh cu˙’a t´ ac gia˙’ liˆ en quan d ¯ˆ e´n luˆ a.n ´ an 96 T` liˆ e.u tham kha˙’o 97 `.NG DUNG ´ KY ´ HIE ˆ U THU.O ` DANH MU C CAC `eu • IRn : Khˆong gian Euclide n chiˆ • · : Chuˆa˙’n Euclide IRn • x, y : T´ıch vˆo hu.o´.ng cu˙’a v´ec to x, y `au mo˙’ b´an k´ınh r tˆam x • B(x, r) := {y | y − x < r} : H`ınh cˆ ¯ r) := {y | y − x ≤ r} : H`ınh cˆ `au d¯o´ng b´an k´ınh r tˆam x • B(x, • A ∈ IRn×n , A : Ma trˆa.n d¯oˆ´i x´ u.ng x´ac d¯.inh du.o.ng • AT : Ma trˆa.n chuyˆe˙’n vi cu˙’a ma trˆa.n A • λmin , (λmax ) : Gi´a tri riˆeng nho˙’ nhˆa´t (l´o.n nhˆa´t) cu˙’a ma trˆa.n A • λ(A) : Tˆa.p c´ac gi´a tri riˆeng cu˙’a ma trˆa.n A √ • A = { max λ | λ ∈ λ(AT A)} : Chuˆa˙’n cu˙’a ma trˆa.n A IRnìn `oi nga.t ã f (x) = Ax, x + b, x : H`am to`an phu.o.ng lˆ • p(x), supx∈D |p(x)| ≤ s v´o.i s ∈ [0, +∞[ : H`am nhiˆ˜eu gi´o.i nˆo.i `oi ngˇa.t v´o.i nhiˆ˜eu gi´o.i nˆo.i • f˜ = f + p : H`am to`an phu.o.ng lˆ • f (x) := Ax, x + b, x → inf, x ∈ D : B`ai to´an quy hoa.ch to`an phu.o.ng (P ) • f (x) := Ax, x + b, x → sup, x ∈ D : B`ai to´an quy hoa.ch to`an phu.o.ng (Q) • f (x) := Ax, x + b, x + p(x) → inf, x ∈ D : B`ai to´an quy hoa.ch `oi ngˇa.t v´o.i nhiˆ˜eu (P˜ ) to`an phu.o.ng lˆ • f (x) := Ax, x + b, x + p(x) → sup, x ∈ D : B`ai to´an quy hoa.ch ˜ `oi ngˇa.t v´o.i nhiˆ˜eu (Q) to`an phu.o.ng lˆ • ∂g(x∗ ) : Du.o´.i vi phˆan cu˙’a g ta.i d¯iˆe˙’m x∗ • L(x, µ0 , , µm ) := m i=0 ài gi (x) : H`am Lagrange ã Tnh chˆa´t (Mγ ) : Mˆo˜i d¯iˆe˙’m γ-cu c tiˆe˙’u x∗ cu˙’a f l`a d¯iˆe˙’m cu c tiˆe˙’u to`an cu.c • T´ınh chˆa´t (Iγ ) : Mˆo˜i d¯iˆe˙’m γ-infimum x∗ cu˙’a f l`a d¯iˆe˙’m infimum to`an cu.c • Lα (f˜) := {x | x ∈ D, f˜(x) ≤ α}, α ∈ IR : Tˆa.p m´ u.c du.o´.i cu˙’a h`am f˜ = f + p (f (x0 ) • h1 (γ) := inf x0 , x1 ∈D, x0 −x1 =γ • h2 (γ) := inf x0 , x1 ∈D, x0 −x1 =γ,−x0 +2x1 ∈D + f (x1 )) − f ( 12 (x0 + x1 )) f (x0 )−2f (x1 )+f (−x0 +2x1 ) • aff D : Bao aphin cu˙’a tˆa.p D `oi d¯a diˆe.n D • ext D : Tˆa.p c´ac d¯iˆe˙’m cu c biˆen cu˙’a tˆa.p lˆ • JD (x∗ ) := ext D \ {x∗ }, x∗ ∈ ext D • d(x, D) := inf y∈D x − y : Khoa˙’ng c´ach t` u x d¯ˆe´n D `oi cu˙’a tˆa.p D • conv D : Bao lˆ • dD := minx∗ ∈ext D {d x∗ , conv JD (x∗ ) } • D(x∗ , β) := {x ∈ D | x = (1 − α)x∗ + αy, y ∈ D, ≤ α ≤ − β}, x∗ ∈ ext D, β ∈ [0, 1] • C (D) := {p : D → IR | p C0 := supx∈D |p(x)| < +∞} ¯C (0, r) : H`ınh cˆ `au d¯o´ng b´an k´ınh r tˆam C (D) • B ˙’ D ˆU -` MO A `en thˆo´ng c´o da.ng B`ai to´an quy hoa.ch to`an phu.o.ng truyˆ f (x) := Ax, x + b, x → inf, x∈D d¯o´ A ∈ IRn×n l`a ma trˆa.n vuˆong, b ∈ IRn l`a v´ec to v`a D ⊂ IRn l`a tˆa.p `oi lˆ `oi, b`ai to´an quy hoa.ch to`an phu.o.ng C` ung v´o.i b`ai to´an quy hoa.ch lˆ `eu nh`a to´an ho.c Viˆe.t nam v`a quˆo´c tˆe´ nghiˆen c´ d¯u.o c nhiˆ u.u, v´ı du nhu H W Kuhn v`a A W Tucker [22], B Bank v`a R Hasel [5], E Blum v`a W Oettli [7], B C Eaves [12], M Frank v`a P Wolfe [13], O L Magasarian [26], G M Lee, N N Tam v`a N D Yen [31], H X Phu [45], H X Phu v`a N D Yen [53], M Schweighofer [57], H Tuy [63], [64], [72], H H Vui v`a P T Son [66] u.u c´ac b`ai to´an C´ac kˆe´t qua˙’ quan tro.ng d¯a˜ thu d¯u.o c nghiˆen c´ `on ta.i nghiˆe.m tˆo´i `e su tˆ quy hoa.ch to`an phu.o.ng cu˙’a c´ac nh`a to´an ho.c l`a vˆ `eu kiˆe.n cˆ `an tˆo´i u.u, d¯iˆ `eu kiˆe.n d¯u˙’ tˆo´i u.u, thuˆa.t to´an t`ım nghiˆe.m tˆo´i u.u, d¯iˆ u.u, t´ınh oˆ˙’n d¯.inh cu˙’a nghiˆe.m tˆo´i u.u c´ac b`ai to´an trˆen bi t´ac d¯ˆo.ng bo˙’.i `eu kˆe´t qua˙’ nghiˆen c´ `e b`ai to´an trˆen d¯a˜ d¯u.o c u nhiˆ˜eu Nhiˆ u.u vˆ ´.ng du.ng d¯ˆe˙’ gia˙’i c´ac b`ai to´an kinh tˆe´ v`a k˜ y thuˆa.t, nhu b`ai to´an lu a cho.n d¯`ˆau tu (portfolio selection) ([27], [28]), b`ai to´an ph´at d¯iˆe.n tˆo´i u.u (economic power dispatch) ([6], [11], [69]), b`ai to´an kinh tˆe´ d¯oˆ´i s´anh (matching economic), ([17]), b`ai to´an m´ay hˆo˜ tro v´ec to (support vector machine) ([29]) Khi A l`a nu˙’.a x´ac d¯i.nh du.o.ng hoˇa.c nu˙’.a x´ac d¯i.nh ˆam th`ı b`ai to´an trˆen c´o thˆe˙’ phˆan r˜a th`anh hai b`ai to´an kh´ac sau: f (x) := Ax, x + b, x → inf, x∈D (P ) f (x) := Ax, x + b, x → sup, x ∈ D (Q) v`a `oi ngˇa.t Luˆa.n ´an n`ay nghiˆen c´ u.u c´ac b`ai to´an quy hoa.ch to`an phu.o.ng lˆ v´o.i nhiˆ˜eu gi´o.i nˆo.i sau: f˜(x) := Ax, x + b, x + p(x) → inf, x∈D (P˜ ) f˜(x) := Ax, x + b, x + p(x) → sup, x ∈ D, ˜ (Q) v`a `eu kiˆe.n supx∈D |p(x)| ≤ s v´o.i gi´a tri d¯o´ p : D → IR tho˙’a m˜an d¯iˆ ˜ d¯u.o c gia˙’ thiˆe´t l`a s ∈ [0, +∞[ v`a A c´ac b`ai to´an (P ), (Q), (P˜ ) v`a (Q) ma trˆa.n d¯oˆ´i x´ u.ng x´ac d¯.inh du.o.ng V`ı c´ac b`ai to´an trˆen d¯u.o c cho.n d¯ˆe˙’ nghiˆen c´ u.u? R˜o r`ang, s = ˜ ch´ınh l`a c´ac b`ai to´an (P ) v`a (Q), hay n´oi c´ach th`ı c´ac b`ai to´an (P˜ ) v`a (Q) kh´ac c´ac b`ai to´an (P ) v`a (Q) l`a c´ac tru.o`.ng ho p riˆeng cu˙’a c´ac b`ai to´an (P˜ ) ˜ D - aˆy l`a l´ v`a (Q) y d¯ˆe˙’ tiˆe´n h`anh nghiˆen c´ u.u c´ac b`ai to´an trˆen, tˆo´i thiˆe˙’u y thuyˆe´t Tuy nhiˆen, c`on mˆo.t sˆo´ l´ y thu c tˆe´ kh´ac du.o´.i t` u quan d¯iˆe˙’m l´ ˜ l`a thu c su cˆ `an d¯aˆy, cho thˆa´y viˆe.c nghiˆen c´ u.u c´ac b`ai to´an (P˜ ), (Q) L´ y th´ u nhˆa´t: f (x) = Ax, x + b, x l`a h`am mu.c tiˆeu ban d¯`aˆu v`a `om c´ac t´ac d¯oˆ ng bˆo˙’ sung p l`a h`am nhiˆ˜eu n`ao d¯o´ H`am nhiˆ˜eu p c´o thˆe˙’ bao gˆ (tˆa´t d¯.inh hoˇa.c ngˆa˜u nhiˆen) lˆen h`am mu.c tiˆeu v`a c´ac lˆo˜i gˆay qu´a - iˆe˙’m d¯aˇ c biˆe.t l`a o˙’ chˆo˜, ch´ ung tr`ınh mˆo h`ınh h´oa, d¯o d¯a.c, t´ınh to´an D ta ha.n chˆe´ chı˙’ x´et nhiˆ˜eu gi´o.i nˆo.i Ha.n chˆe´ n`ay l`a khˆong qu´a ngˇa.t, c´o thˆe˙’ `eu b`ai to´an thu c tˆe´, chˇa˙’ng ha.n nhu hai v´ı d¯u.o c tho˙’a m˜an nhiˆ du minh ho.a sau d¯ˆay Mˆo.t nh˜ u.ng u ´.ng du.ng nˆo˙’i bˆa.t cu˙’a quy hoa.ch to`an phu.o.ng l`a b`ai to´an lu a cho.n d¯`ˆau tu (H M Markowitz [27], [28]) B`ai to´an ph´at biˆe˙’u nhu sau: Phˆan phˆo´i vˆo´n qua n ch´ u.ng kho´an (asset) c´o sˇa˜n d¯ˆe˙’ c´o thˆe˙’ gia˙’m thiˆe˙’u ru˙’i ro v`a tˆo´i d¯a lo i nhuˆa.n, t´ u.c l`a t`ım v´ec to tı˙’ lˆe x ∈ D, D := {x = (x1 , x2 , , xn ) | nj=1 xj = 1} d¯ˆe˙’ f (x) = ωxT Σx − ρT x d¯a.t gi´a tri nho˙’ nhˆa´t, d¯´o xj , j = 1, , n, l`a ty˙’ lˆe ch´ u.ng kho´an th´ u j danh mu.c d¯`ˆau tu., ω l`a tham sˆo´ ru˙’i ro, Σ ∈ IRn×n l`a ma trˆa.n hiˆe.p phu.o.ng sai, ρ ∈ IRn l`a v´ec to lo i nhuˆa.n k` y vo.ng V`ı Σ v`a ρ thu.o`.ng ˜ v`a ρ˜, d¯o´ ch´ khˆong d¯u.o c x´ac d¯i.nh ch´ınh x´ac m`a chı˙’ xˆa´p xı˙’ bo˙’.i Σ ung ˜ − ρ˜T x = f (x) + p(x), d¯´o ta pha˙’i cu c tiˆe˙’u h´oa h`am f˜(x) = ωxT Σx ˜ − Σ)x − (˜ p(x) = ωxT (Σ ρ − ρ)T x Khi quy d¯.inh, khˆong d¯u.o c b´an khˆo´ng, t´ u.c l`a xj ≥ 0, j = 1, , n, th`ı tˆa.p chˆa´p nhˆa.n d¯u.o c D l`a gi´o.i nˆo.i V`ı vˆa.y nhiˆ˜eu p c˜ ung gi´o.i nˆo.i trˆen D N´oi mˆo.t c´ach tˆo˙’ng qu´at, t´ınh gi´o.i nˆo.i cu˙’a nhiˆ˜eu luˆon d¯u.o c d¯a˙’m ba˙’o D gi´o.i nˆo.i v`a p liˆen tu.c trˆen D Gia˙’ thiˆe´t `eu b`ai to´an thu c tˆe´ n`ay c˜ ung ph` u ho p v´o.i nhiˆ Mˆo.t v´ı du n˜ u.a cho thˆa´y l`a nhiˆ˜eu gi´o.i nˆo.i luˆon xuˆa´t hiˆe.n gia˙’i mˆo.t `an l´o.n c´ac sˆo´ b`ai to´an tˆo´i u.u (P ) hoˇa.c (Q) n`ao d¯´o bˇa` ng m´ay t´ınh Do phˆ `au hˆe´t thu c khˆong thˆe˙’ biˆe˙’u diˆ˜en ch´ınh x´ac bˇa` ng m´ay t´ınh, nˆen d¯ˆo´i v´o.i hˆ x ∈ D ta khˆong thˆe˙’ t´ınh ch´ınh x´ac d¯a.i lu.o ng f (x) = Ax, x + b, x m`a chı˙’ c´o thˆe˙’ xˆa´p xı˙’ f (x) bo˙’.i mˆo.t sˆo´ dˆa´u chˆa´m d¯oˆ ng f˜(x) n`ao d¯´o H`am f˜ `oi, khˆong to`an phu.o.ng v`a thˆa.m ch´ı l`a khˆong liˆen tu.c trˆen D Khi khˆong lˆ d¯o´ h`am p := f˜− f mˆo ta˙’ c´ac lˆo˜i t´ınh to´an C´ac lˆo˜i d¯´o bi chˇa.n bo˙’.i mˆo.t cˆa.n trˆen s ∈ [0, +∞[ n`ao d¯o´ c´o thˆe˙’ u.o´.c lu.o ng d¯u.o c, t´ u.c l`a supx∈D |p(x)| ≤ s Ngo`ai ra, bˇa` ng c´ach su˙’ du.ng c´ac sˆo´ dˆa´u chˆa´m d¯oˆ ng d`ai ho.n v`a/hoˇa.c c´ac thuˆa.t to´an tˆo´t ho.n, ta c´o thˆe˙’ gia˙’m cˆa.n trˆen s L´ y th´ u hai: f˜ l`a h`am mu.c tiˆeu d¯´ıch thu c v`a f l`a h`am mu.c tiˆeu `eu d¯u.o c l´ y tu.o˙’.ng h´oa hoˇa.c l`a h`am mu.c tiˆeu thay thˆe´ Trong thu c tˆe´, nhiˆ `oi, hoˇa.c to`an h`am thˆe˙’ hiˆe.n mˆo.t sˆo´ mu.c tiˆeu thu c tiˆ˜en d¯u.o c gia˙’ d¯.inh l`a lˆ phu.o.ng, hoˇa.c c´o mˆo.t sˆo´ t´ınh chˆa´t thuˆa.n tiˆe.n d¯a˜ d¯u.o c nghiˆen c´ u.u k˜ y, hoˇa.c - iˆ `eu n`ay d¯˜a d¯u.o c dˆ˜e nghiˆen c´ u.u, nhu.ng thu c th`ı khˆong pha˙’i l`a nhu vˆa.y D H X Phu, H G Bock v`a S Pickenhain d¯`ˆe cˆa.p d¯ˆe´n [48] Trong bˆo´i ca˙’nh d¯´o, p = f˜ − f l`a h`am hiˆe.u chı˙’nh C´o thˆe˙’ gia˙’ thiˆe´t p l`a gi´o.i nˆo.i (tˆo´i thiˆe˙’u trˆen tˆa.p chˆa´p nhˆa.n d¯u.o c) bo˙’.i mˆo.t sˆo´ du.o.ng kh´a b´e s, v`ı nˆe´u |p(x)| qu´a l´o.n th`ı su thay thˆe´ khˆong c`on ph` u ho p n˜ u.a - ˆe˙’ gia˙’i th´ıch d¯iˆ `eu n`ay, ta d¯`ˆe cˆa.p d¯ˆe´n vˆa´n d¯`ˆe thu.o`.ng d¯u.o c nghiˆen c´ D u.u cu˙’a ph´at d¯iˆe.n tˆo´i u.u, t´ u.c l`a b`ai to´an phˆan bˆo´ lu.o ng d¯iˆe.n nˇang cho t` u.ng tˆo˙’ m´ay ph´at nhiˆe.t d¯iˆe.n cho tˆo˙’ng chi ph´ı (gi´a th`anh) l`a cu c tiˆe˙’u, d¯`oˆng `au lu.o ng d¯iˆe.n nˇang v`a thoa˙’ m˜an r`ang buˆo.c th`o.i vˆa˜n d¯a´p u ´.ng d¯u.o c nhu cˆ 90 Ch´ u.ng minh Nˆe´u s = th`ı ξ(s) = nˆen kˆe´t luˆa.n cu˙’a bˆo˙’ d¯`ˆe l`a hiˆe˙’n nhiˆen Nˆe´u s > 0, lˆa´y bˆa´t k` y x∗ ∈ Slocal (0) V´o.i mo.i x ∈ D, x = x∗ ta c´o x − x∗ f (x) = f (x ) + 2Ax + b, x − x∗ nˆen, v´o.i mo.i x ∈ D, x = x∗ th`ı ∗ x − x∗ + A(x − x∗ ), x − x∗ , ∗ x − x∗ f (x) = 2Ax + b, x − x∗ x − x∗ + A(x − x∗ ), x − x∗ + 3s + f (x∗ ) − 3s ∗ (4.4.27) X´et biˆe˙’u th´ u.c 2Ax∗ + b, x − x∗ x − x∗ x − x∗ + A(x − x∗ ), x − x∗ + 3s V`ı A(x − x∗ ), x − x∗ ≤ λmax x − x∗ tho˙’a m˜an x − x∗ = ξ(s), ta d¯u.o c v`a ξ(s) > 0, nˆen v´o.i mo.i x ∈ D x − x∗ x − x∗ + A(x − x∗ ), x − x∗ + 3s ∗ x−x ≤ λmax ξ(s) + η0 ξ(s) + 3s 2Ax∗ + b, V`ı s ≤ s0 v`a ξ(s) ≤ ξ(s0 ) nˆen λmax ξ(s)2 + η0 ξ(s) + 3s ≤ λmax ξ(s0 )2 + η0 ξ(s0 ) + 3s0 Thay s0 , ξ(s) d¯u.o c x´ac d¯.inh theo (4.4.25) v`a (4.4.26) ta suy λmax ξ(s0 )2 + η0 ξ(s0 ) + 3s0 = Do d¯o´ x − x∗ 2Ax + b, x − x∗ ∗ x − x∗ + A(x − x∗ ), x − x∗ + 3s ≤ Kˆe´t ho p biˆe˙’u th´ u.c n`ay v´o.i (4.4.27) ta suy kˆe´t luˆa.n cu˙’a bˆo˙’ d¯`ˆe 91 ˜ Khi d¯´o - i.nh l´ D y 4.4.25 X´et B` to´an (Q) ¯C (0, s0 ) : ∀p ∈ B max x∗ ∈Slocal (0) d x∗ , Slocal (p) ≤ ξ( p C ) `eu khˇa˙’ng d¯.inh l`a hiˆe˙’n nhiˆen Ch´ u.ng minh Nˆe´u p ≡ th`ı d¯iˆ - aˇ t s := p C Do D l`a Lˆa´y bˆa´t k` y p ∈ C (D) cho < p C ≤ s0 D ¯ x∗ , ξ(s) ∩D l`a tˆa.p compact V`ı h`am to`an phu.o.ng lˆ `oi `oi d¯a diˆe.n nˆen B tˆa.p lˆ ¯ x∗ , ξ(s) ∩D, nˆen ngˇa.t bi nhiˆ˜eu gi´o.i nˆo.i f˜ = f +p bi chˇa.n (trˆen) trˆen tˆa.p B ¯ x∗ , ξ(s) ∩ D `on ta.i (˜ supx∈B(x a.t kh´ac, tˆ xi ) B ¯ ∗ , ξ(s))∩D f˜(x) < +∞ Mˇ cho limi→∞ x˜i = x˜∗ v`a limi→∞ f˜(˜ xi ) = supx∈B(x ¯ ∗ , ξ(s))∩D f˜(x) Ta khˇa˙’ng d¯i.nh x˜∗ ∈ B(x∗ , ξ(s)) ∩ D Thˆa.t vˆa.y gia˙’ su˙’ kˆe´t luˆa.n trˆen l`a ¯ x∗ , ξ(s) \ B(x∗ , ξ(s) ∩ D Ta c´o sai, d¯´o x˜∗ ∈ B sup ¯ ∗ , ξ(s))∩D x∈B(x f˜(x) = lim f˜(˜ xi ) i→∞ = lim f (˜ xi ) + p(˜ xi ) i→∞ (4.4.28) `on ta.i limi→∞ f˜(˜ `oi V`ı tˆ xi ) = limi→∞ f (˜ xi ) + p(˜ xi ) , m`a h`am to`an phu.o.ng lˆ `on ta.i limi→∞ p(˜ ngˇa.t f liˆen tu.c ta.i mo.i d¯iˆe˙’m thuˆo.c D nˆen tˆ xi ) Do d¯o´ thay limi→∞ f˜(˜ xi ) + p(˜ xi ) = limi→∞ f (˜ xi ) + limi→∞ p(˜ xi ) v`ao (4.4.28) v`a ch´ u y ´ rˇ`a ng f (˜ x∗ ) ≤ f (x∗ ) − 3s theo Bˆo˙’ d¯`ˆe 4.4.8, ta d¯u.o c sup ¯ ∗ , ξ(s))∩D x∈B(x f˜(x) = lim f (˜ xi ) + lim p(˜ xi ) i→∞ ∗ i→∞ ≤ f (˜ x )+s ≤ f (x∗ ) − 3s + s ≤ f (x∗ ) + p(x∗ ) − s = f˜(x∗ ) − s Suy sup ¯ ∗ , ξ(s))∩D x∈B(x f˜(x) ≤ sup f˜(x) − s ¯ ∗ , ξ(s))∩D x∈B(x Biˆe˙’u th´ u.c nhˆa.n d¯u.o c l`a vˆo l´ y, nˆen x˜∗ ∈ B(x∗ , ξ(s)) ∩ D, d¯o´ suy x˜∗ ∈ D l`a d¯iˆe˙’m supremum d¯i.a phu.o.ng cu˙’a f˜ = f + p trˆen D Ngo`ai v`ı 92 ¯ ∗ , ξ(s)) nˆen x˜∗ ∈ B(x max x∗ ∈Slocal (0) d x∗ , Slocal (p) = ≤ ≤ = max x∗ ∈Slocal (0) max x∗ ∈Slocal (0) inf y˜∗ ∈Slocal (p) x˜∗ − x∗ max ξ(s) max ξ( p x∗ ∈Slocal (0) x∗ ∈Slocal (0) y˜∗ − x∗ C ) T´om la.i ta nhˆa.n d¯u.o c ¯C (0, s0 ) : ∀p ∈ B max x∗ ∈Slocal (0) d x∗ , Slocal (p) ≤ ξ( p C ) - i.nh l´ D y d¯a˜ d¯u.o c ch´ u.ng minh `e 4.4.30 H` Mˆ e.nh d ¯ˆ am d¯a tri Slocal (p) l`a nu˙’.a liˆen tu.c du.´o.i ta.i d¯iˆe˙’m Ch´ u.ng minh Theo d¯.inh ngh˜ıa h`am nu˙’.a liˆen tu.c du.o´.i, nˆe´u Slocal (0) = ∅ th`ı Slocal (p) nu˙’.a liˆen tu.c du.o´.i ta.i l`a hiˆe˙’n nhiˆen Ta x´et tru.o`.ng ho p Slocal (0) = ∅ Lˆa´y tˆa.p mo˙’ bˆa´t k` y V ⊂ IRn tho˙’a m˜an Slocal (0) ∩ V = ∅ Khi d¯o´ ∃ x∗ ∈ Slocal (0) : x∗ ∈ V ¯ ∗ , δ) ⊂ V Mˇa.t kh´ac, `on ta.i δ > cho B(x V`ı V l`a tˆa.p mo˙’., ta suy tˆ ξ(s) d¯u.o c x´ac d¯.inh bo˙’.i biˆe˙’u th´ u.c (4.4.26) nˆen lim ξ(s) = lim − η0 − s→0 s→0 = η02 − 12λmax s /(2λmax ) − η0 − (−η0 ) /(2λmax ) = Biˆe˙’u th´ u.c n`ay cho ph´ep ta cho.n sˆo´ du.o.ng s1 ≤ s0 cho v´o.i mo.i s ∈ ]0, s1 ] th`ı ξ(s) ≤ δ, v`ı vˆa.y ¯ x∗ , ξ(s) ∩ D ⊂ B(x ¯ ∗ , δ) ∩ D ⊂ V B (4.4.29) - i.nh l´ Mˇa.t kh´ac, theo D y 4.4.25, v´o.i mo.i p ∈ C (D) tho˙’a m˜an p C ≤ s0 ¯ x∗ , ξ( p C ) ∩ D l`a d¯iˆe˙’m `on ta.i x˜∗ ∈ B (d¯u.o c x´ac d¯i.nh theo (4.4.25)) tˆ 93 supremum d¯.ia phu.o.ng cu˙’a f˜ = f + p, t´ u.c l`a x˜∗ ∈ Slocal (p) V`ı s1 ≤ s0 nˆen kˆe´t ho p v´o.i biˆe˙’u th´ u.c (4.4.29) ta nhˆa.n d¯u.o c ¯C (0, s1 ) : V ∩ Slocal (p) = ∅, ∀p ∈ B Do d¯o´ Slocal (p) l`a nu˙’.a liˆen tu.c du.o´.i ta.i d¯iˆe˙’m T` u d¯.inh ngh˜ıa 4.3.14 ta suy h`am Slocal (p) khˆong nu˙’.a liˆen tu.c trˆen `on ta.i lˆan cˆa.n V tho˙’a m˜an Slocal (0) ⊆ V v`a d˜ay (pi ), i = 1, ta.i nˆe´u tˆ `e C (D) cho Sglobal (pi ) \ V = ∅ v´o.i mo.i i = 1, 2, hˆo.i tu vˆ V´ı du sau d¯aˆy chı˙’ Slocal (p) khˆong nu˙’.a liˆen tu.c trˆen ta.i V´ı du 4.4.14 Cho f (x) = x2 , x ∈ [ 0, ] 1/i − 2x2 nˆe´u x ∈ [ 0, 1/i ] pi (x) = nˆe´u x ∈ [ 0, ] \ [ 0, 1/i ] i = 1, Ta t´ınh d¯u.o c pi 0C (D) = 1/i, Slocal (0)={2} v`a Slocal (pi ) = {0, 2}, i = 1, 2, Lˆa´y tˆa.p mo˙’ V = ]1.5, 2.1[ ta c´o {2} = Slocal (0) ⊂ V = ]1.5, 2.1[ Trong d¯o´ v´o.i mo.i i th`ı ∈ Slocal (pi ) nhu.ng ∈ / V, nˆen suy Slocal (pi ) \ V = ∅ Do d¯o´ Slocal (p) khˆong nu˙’.a liˆen tu.c trˆen ta.i Kˆ e´t luˆ a.n: C´ac kˆe´t qua˙’ d¯a.t d¯u.o c, co ba˙’n d¯u.o c tr`ınh b`ay c´ac Mu.c `om: mˆo.t sˆo´ d¯iˆ `eu kiˆe.n d¯u˙’ d¯ˆe˙’ h`am to`an phu.o.ng lˆ `oi 4.1–4.4, ch´ ung bao gˆ `oi (Mˆe.nh d¯`ˆe 4.1.24); c´ac t´ınh chˆa´t cu˙’a ngˇa.t bi nhiˆ˜eu gi´o.i nˆo.i l`a γ-lˆ ˜ (c´ac mˆe.nh d¯`ˆe c´ac d¯iˆe˙’m cu c d¯a.i v`a supremum to`an cu.c cu˙’a B`ai to´an (Q) 4.2.25, 4.2.26, 4.2.27, 4.2.28); t´ınh ˆo˙’n d¯i.nh, nu˙’.a liˆen tu.c trˆen cu˙’a h`am tˆa.p ˜ (D - i.nh l´ c´ac d¯iˆe˙’m supremum to`an cu.c cu˙’a B`ai to´an (Q) y 4.3.24, Mˆe.nh d¯`ˆe 4.3.29); t´ınh oˆ˙’n d¯.inh, nu˙’.a liˆen tu.c du.o´.i cu˙’a h`am tˆa.p c´ac d¯iˆe˙’m supremum ˜ (D - i.nh l´ d¯.ia phu.o.ng cu˙’a B`ai to´an (Q) y 4.4.25, Mˆe.nh d¯`ˆe 4.4.30) 94 ˆ´T LUA ˆ N CHUNG KE `e: ac vˆ a´n d ¯ˆ Luˆ a.n ´ an d ¯˜ a gia˙’i quyˆ e´t d ¯u.o c c´ `oi ngo`ai v´o.i mo.i γ ≥ γ ∗ , d¯o´ • Chı˙’ h`am bi nhiˆ˜eu f˜ = f + p l`a γ-lˆ γ ∗ = 2s/λmin ; d¯iˆe˙’m γ ∗ -cu c tiˆe˙’u cu˙’a f˜ l`a d¯iˆe˙’m cu c tiˆe˙’u to`an cu.c; d¯u.o`.ng k´ınh cu˙’a tˆa.p c´ac d¯iˆe˙’m cu c tiˆe˙’u to`an cu.c cu˙’a B`ai to´an (P˜ ) nho˙’ ho.n hoˇa.c bˇa` ng γ ∗ ; khoa˙’ng c´ach gi˜ u.a d¯iˆe˙’m cu c tiˆe˙’u to`an cu.c cu˙’a B`ai to´an (P˜ ) v`a d¯iˆe˙’m cu c tiˆe˙’u to`an cu.c cu˙’a h`am f nho˙’ ho.n hoˇa.c bˇa` ng `eu kiˆe.n tˆo´i u.u suy rˆo.ng γ ∗ Ngo`ai t´ınh chˆa´t tu a thˆo v`a mˆo.t sˆo´ d¯iˆ cu˙’a h`am f˜ c˜ ung d¯u.o c tr`ınh b`ay C´ac kˆe´t qua˙’ trˆen d¯˜a d¯u.o c cˆong bˆo´ b`ai b´ao “Global infimum of strictly convex quadratic functions with bounded perturbation” (xem Danh mu.c c´ac cˆong tr`ınh cu˙’a t´ac gia˙’ liˆen quan d¯ˆe´n luˆa.n ´an) `oi ngo`ai v´o.i tˆa.p cˆan d¯aˇ c biˆe.t Γ ⊂ IRn ; • Ch´ u.ng minh d¯u.o c, h`am f˜ l`a Γ-lˆ d¯iˆe˙’m Γ-tˆo´i u.u d¯.ia phu.o.ng cu˙’a B`ai to´an (P˜ ) l`a d¯iˆe˙’m tˆo´i u.u to`an cu.c; hiˆe.u cu˙’a hai nghiˆe.m tˆo´i u.u bˆa´t k` y cu˙’a B`ai to´an (P˜ ) n`am tˆa.p Γ; x∗ − x˜∗ ∈ 12 Γ nˆe´u x∗ l`a nghiˆe.m cu c tiˆe˙’u to`an cu.c cu˙’a f trˆen D v`a x˜∗ l`a nghiˆe.m tˆo´i u.u to`an cu.c bˆa´t k` y cu˙’a B`ai to´an (P˜ ); tˆa.p nghiˆe.m tˆo´i - i.nh u.u Ss cu˙’a (P˜ ) l`a oˆ˙’n d¯i.nh theo khoa˙’ng c´ach Hausdorff dH (.,.) D l´ y Kuhn-Tucker suy rˆo.ng cho B`ai to´an (P˜ ) c˜ ung d¯u.o c ch´ u.ng minh C´ac kˆe´t qua˙’ trˆen d¯˜a d¯u.o c d¯ˇang ta˙’i b`ai b´ao “ Some properties of boundedly disturbed strictly convex quadratic functions” (xem Danh mu.c c´ac cˆong tr`ınh cu˙’a t´ac gia˙’ liˆen quan d¯ˆe´n luˆa.n ´an) `oi v´o.i γ ≥ (2/λmin ) v`a γ-lˆ `oi ngˇa.t v´o.i • Chı˙’ h`am f˜ l`a γ-lˆ 1 γ > (2/λmin ) ; D bi chˇa.n v`a γ = (2/λmin ) , mo.i d¯iˆe˙’m supremum ˜ chı˙’ c´o thˆe˙’ l`a d¯iˆe˙’m γ- cu c biˆen cu˙’a D v`a c´o to`an cu.c cu˙’a B`ai to´an (Q) ´ıt nhˆa´t mˆo.t d¯iˆe˙’m l`a γ-cu c biˆen ngˇa.t Mˆo.t sˆo´ t´ınh chˆa´t quan tro.ng cu˙’a tˆa.p c´ac d¯iˆe˙’m supremum to`an cu.c Sglobal (p) v`a tˆa.p c´ac d¯iˆe˙’m supremum ˜ nhu t´ınh oˆ˙’n d¯i.nh v`a t´ınh nu˙’.a d¯.ia phu.o.ng Slocal (p) cu˙’a B`ai to´an Q 95 `an l´o.n c´ac kˆe´t qua˙’ d¯u.o c liˆe.t kˆe o˙’ trˆen liˆen tu.c c˜ ung d¯u.o c chı˙’ Phˆ d¯a˜ d¯u.o c cˆong bˆo´ b`ai b´ao “Maximizing strictly convex quadratic functions with bounded perturbation” (xem Danh mu.c c´ac cˆong tr`ınh cu˙’a t´ac gia˙’ liˆen quan d¯ˆe´n luˆa.n ´an) `an tiˆ a´n cˆ e´p tu.c nghiˆ en c´ u.u: Nh˜ u.ng vˆ `e l´ Luˆa.n ´an chı˙’ m´o.i d¯`ˆe cˆa.p d¯ˆe´n mˆo.t sˆo´ vˆa´n d¯`ˆe vˆ y thuyˆe´t cu˙’a B`ai to´an `oi ngˇa.t v´o.i nhiˆ˜eu gi´o.i nˆo.i Do d¯´o ch´ quy hoa.ch to`an phu.o.ng lˆ ung tˆoi c`on u.ng vˆa´n d¯`ˆe sau d¯aˆy tiˆe´p tu.c nghiˆen c´ u.u nh˜ • Xˆay du ng thuˆa.t to´an t´ınh to´an t`ım l`o.i gia˙’i tˆo´i u.u cu˙’a c´ac b`ai to´an ˜ (P˜ ) v`a (Q) ´ du.ng thuˆa.t to´an t´ınh to´an t`ım l`o.i gia˙’i tˆo´i u.u cu˙’a c´ac b`ai to´an (P˜ ) • Ap ˜ v`ao c´ac b`ai to´an thu c tˆe´ nhu b`ai to´an ph´at d¯iˆe.n tˆo´i u.u, kinh v`a (Q) tˆe´ d¯oˆ´i s´anh, Tuy nhiˆen, v`ı th`o.i gian ha.n he.p nˆen ch´ ung tˆoi c˜ ung chu.a tra˙’ l`o.i d¯u.o c c´ac vˆa´n d¯`ˆe trˆen Ch´ ung tˆoi hy vo.ng rˇ`a ng c´ac vˆa´n d¯`ˆe n`ay s˜e s´o.m d¯u.o c gia˙’i quyˆe´t 96 ˆ DANH MU TR`INH C CONG ´ GIA˙’ LIEN ˆ QUAN D ˆ´N LUA ˆ N AN ´ -E CU˙’A TAC C´ac b`ai b´ao cu˙’a t´ac gia˙’ liˆen quan d¯ˆe´n luˆa.n ´an l`a: H X Phu and V M Pho, Global infimum of strictly convex quadratic functions with bounded perturbation, Mathematical Methods of Operations Research, 72(2), 2010, 327–345 H X Phu and V M Pho, Some properties of boundedly disturbed strictly convex quadratic functions, Optimization, DOI 10.1080/0233193100746114, Published online: 07 May 2010 H X Phu, V M Pho and P T An, Maximizing strictly convex quadratic functions with bounded perturbation, Journal of Optimization Theory and Applications, 149(1) 2011, 1–25 ˙’ O ` LIE ˆ U THAM KHA TAI `oi thˆ `oi suy rˆo.ng v´o.i nhiˆ˜eu [1] P T An, H` am lˆ o v`a t´ınh ˆo˙’n d¯.inh cu˙’ a h`am lˆ - a.i ho.c Vinh, 1999 tuyˆe´n t´ınh, Luˆa.n ´an Tiˆe´n s˜ı To´an ho.c, D - a.i ho.c Quˆo´c gia, H`a nˆo.i, [2] Pha.m K` y Anh, Gia˙’ i t´ıch sˆo´, Nh`a xuˆa´t ba˙’n D 1998 `oi v`a γ-du.´o.i vi phˆan, [3] N N Hai, Mˆ o.t sˆ o´ t´ınh chˆa´t gia˙’ i t´ıch cu˙’ a h`am γ-lˆ Luˆa.n ´an Tiˆe´n s˜ı To´an ho.c, H`a nˆo.i, 2001 [4] J P Aubin and H Frankowska, Set-Valued Analysis, Birkhăauser, Boston Basel Berlin, 1990 [5] B Bank and R Hasel, Stability of mixed-integer quadratic programming problems, Mathematical Programming Study, 21, 1984, 1–17 [6] P P J van den Bosch and F A Lootsma, Scheduling of power generation via large-scale nonliear optimization, Journal of Optimization Theory and Applications, 55, 1987, 313–326 [7] E Blum and W Oettli, Direct proof of the existence theorem for quadratic programming, Operations Research, 20, 1973, 165–167 [8] M J Canovas, A Hantoute, M Lopez and A Marco, Lipschitz behavior of convex semi-infinite optimization problems: a variational approach, Journal of Global Optimization, 41 2008, 1–13 [9] C H Chen and S N Yeh, Particle swarm optimization for economic power dispatch with valve- point effects, Transmission and Distribution 97 98 Conference and Exposition Latin America, Venezuela, 1–4244–0288– 3/20.00, 2006, IEEE [10] J W Daniel, Stability of the solution of definite quadratic programs, Mathematical Programming, 5, 1973, 41–53 [11] R M S Danaraj and F Gajendran, Quadratic programming solution to emission and economic dispatch problem, IE(I) Journal-EL, 86, 2005, 129–132 [12] B C Eaves, On quadratic programming, Management Science, 17, 1971, 698–711 [13] M Frank and P Wolfe, An algorithm for quadratic programming, Naval Research Logistics Quarterly, 3, 1956, 95–110 [14] M A Hanson, On sufficiency of the Kuhn-Tucker conditions, Journal of Mathematical Annalysis and Applications, 80, 1981, 545–550 [15] H Hartwig, Local boundedness and continuity of generalized convex functions, Optimization, 26, 1992, 1–13 [16] T C Hu, V Klee and D Larman, Optimization of globally convex functions, SIAM Journal Control Optimization, 27, 1989, 1026–1047 [17] T R Jefferson and C H Scott, Quadratic geometric programming with application to maching economics, Mathematical Programming, 37, 1985,137-152 [18] S Karamardian, Duality in Mathematical Programming, Doctoral Thesis, University of California Press, Berkeley, 1965 [19] J Kennedy and R C Eberhart, Swarm Intelligence, Morgan Kaufimann, 2001 [20] D Klatte, On the lower semicontinuity of optimal sets in convex parametric optimization Point-to-set maps and mathematical programming, Mathematical Programming Study, 10, 1979, 104–109 99 [21] D Klatte, Lower semicontinuity of the minimum in parametric convex programs, Journal of Optimization Theory and Applications 94, 1997, 511–517 [22] H.W Kuhn and A W Tucker, Nonlinear Programming, Proceedings of the Second Berkeley Symposium on the Mathematical Statistics and Probability, University of California Press, Berkeley, 418–492, 1951 [23] B Kummer, Stability of generalized equations and Kuhn-Tucker points of perturbed convex programs, System modelling and optimization (Copenhagen, 1983), 213–218, Lecture Notes in Control and Information Sciences, 59, Springer, Berlin, 1984 [24] O L Mangasarian, Locally unique solutions of quadratic programs, linear and nonlinear complementarity problems, Mathematical Programming, 19, 1980, 200–212 [25] O L Mangasarian, Pseudo-convex functions, SIAM Journal on Control, 3, 1965, 200–212 [26] B Matos, The direct power of adjacent vertex programming methods, Management science, 12, 1965, 241–255 [27] M H Markowitz, Portfolio selectio, Journal Finance, 7, 1952, 77–91 [28] M H Markowitz, Portfolio selectio, Wiley, New York, 1959 [29] D Meyer, F Leisch and K Hornik, The support vector machine under test, Neurocomputing , bf 55, 2003, 169-186 [30] K Mirnia and A Ghaffari-Hadigheh, Support set expansion sensitivity analysis in convex quadratic optimization, Optimization Methods Software, 22, 2007, 601–616 [31] G M Lee, N N Tam and N D Yen, Quadratic Programming and Affine Variational Inequalities, Springer, 2005 100 [32] S H Ling, H K Lam, F H F Leung and Y S Lee, Improved genetic algorithm for economic load dispatch with valve-point loadings, 0-78037906-3/03/$17.00, 2003, IEEE [33] J B Park, Y W Jeong, H H Kim and J R Shin, An Improved particle swarm optimization for economic dispatch with valve-point effect, Journal of Innovations in Energy Systems and Power, 1, 2006, 1–7 [34] H X Phu, γ-subdifferential and γ-convexity of functions on the real line, Applied Mathematics and Optimization, 27, 1993, 145–160 [35] H X Phu, Representation of bounded convex sets by rational convex hull of its gamma-extreme points, Numerical Functional Analysis and Optimization, 19, 1994, 915–920 [36] H X Phu, γ-subdifferential and γ-convex functions on normed spaces, Journal of Optimization Theory and Applications, 85, 1995, 649–676 [37] H X Phu, Some properties of globally δ-convex functions, Optimization, 35, 1995, 22–41 [38] H X Phu, Six kinds of roughly convex functions, Journal of Optimization Theory and Applications, 92, 1997, 357–375 [39] H X Phu, Strictly and roughly convexlike fuctions, Journal of Optimization Theory and Applications, 117, 2003, 139–156 [40] H X Phu, Some basic ideas of rough analysis, Proceeding of the sixth Vietnamese Mathematical Conference, Hue, 9-2002, 3–31, Hanoi National University Publishing House, Hanoi, 2005 [41] H X Phu, Inner γ-convex functions in normed spaces, E-Preprint 2007/01/01, Hanoi Institute of Mathematics, 2007 [42] H X Phu, Outer γ-convexity and inner γ-convexity of disturbed functions, Vietnam Journal of Mathematics, 35, 2007, 107–119 101 [43] H X Phu, Supremizers of inner γ-convex functions, Mathematical Methods of Operations Research, 92(2), 2008, 207–222 [44] H X Phu, Outer Γ-convexity in vector spaces, Numerical Functional Analysis and Optimization, 25(7, 8), 2008, 835–854 [45] H X Phu, Some properties of solution sets to nonconvex quadratic programming problems, Optimization, 56, 2007, 369–383 [46] H X Phu and P T An, Stable generalization of convex functions, Optimization, 38, 1996, 309–318 [47] H X Phu and P T An, Outer γ-convexity in normed linear spaces, Vietnam Journal of Mathematics, 27, 1999, 323–334 [48] H X Phu, G Bock and S Pickenhain, Rough stability of solutions to nonconvex optimization problems, Optimization, Dynamics, and Economic Analysis, 22–35, Physica-Verlag, Heidelberg New York, 2000 [49] H X Phu and N N Hai, Some analytical properties of γ–convex functions on the real line, Journal of Optimization Theory and Applications, 91, 1996, 671–694 [50] H X Phu and V M Pho, Some properties of boundedly disturbed strictly convex quadratic functions, Optimization, DOI 10.1080/0233193100746114, Published online: 07 May 2010 [51] H X Phu and V M Pho, Global infimum of strictly convex quadratic functions with bounded perturbation, Mathematical Methods of Operations Research, 72(2), 2010, 327–345 [52] H X Phu, V M Pho and P T An, Maximizing strictly convex quadratic functions with bounded perturbation, Journal of Optimization Theory and Applications, 149(2), 2011, 1–25 102 [53] H X Phu and N D Yen, On the stability of solutions to quadratic programming problems, Mathematical Programming Study, 89, 2001, 385–394 [54] R T Rockafeller, Convex Analysis, Princeton University Press, Princeton, 1970 [55] R Schultz, Subdifferentials of perturbed convex functions, 16th annual conference on mathematical optimization (Sellin, 1984), 106–113, Seminarberichte, 64, Humboldt University, Berlin [56] R Schultz, Estimates for Kuhn-Tucker points of perturbed convex programs, Optimization, 19, 1988, 29–43 [57] M Schweighofer, Global optimization of polynomiald using gradient tentacles and sums of squares, SIAM Journal Optimization, 106(3), 2006, 920–942 [58] I Singer, Duality theorems for perturbed convex optimization, Journal of Mathematical Analysis and Applications, 81, 1981, 437–452 [59] M Slater, Lagrange multipliers revisited: A contribution to nonlinear programming, Cowles Commission Discussion Paper, Mathematics, 403, 1951 [60] P Suburaj, L Ganesan, K Ramar and Rajkumar, A New approach to economic dispatch problem with valve point loading using evolutionary programming, Interational Journal of Electrical and Power Engineering, 1(2), 2007, 251–254 [61] S S Subramani and P R Rajeswari, A Modified particle swarm optimization for economic dispatch problems with non-smooth cost functions, Interational Journal of Software Computing 3, 2008, 326– 332 103 [62] J I Al-Sumait, A.K Al-Othman and J K Sykulski, Application of pattern method to power system valve-point economic load dispatch, Electrical Power and Energy Systems, 29, 2007, 720–730 [63] H Tuy, On outer approximation methods for solving concave minimization problems, Acta Mathematica Vietnamica 8, 1983, 3–34 [64] H Tuy and N T Hoai-Phuong, A robust algorithm for quadratic optimization under quadratic constraints, Journal Global Optimization, 37, 2007, 527–554 [65] L I Trudzik, Perturbed convex programming in reflexive Banach spaces, Nonliear Analysis, 9, 1985, 61–78 [66] H H Vui and P T Son, Global optimization of polynomiald using the truncated tangency variety and sums of squares, SIAM Journal Optimization, 19(2), 2008, 941–951 [67] D N Wilke, S Kokand and A A Groenwold, Comparision of linear and classical velocity update rules in particle swarm optomization: Notes on diversity, Interational Journal for Numerical Methods in Engineering, 70, 2007, 962–984 [68] E Zeidler, Nonlinear Functional Analysis–Main Principles and Their Applications, Springer-Verlag, New York, 1995 [69] Y Zhu and K Tomsovic, Optimal distribution power flow for systems with disturbed enegy resources, Int J Electr Power Enegy Systems, 29, 2007, 260–267 [70] S Zlobec, Characterizing an optimal input in perturbed convex programming, Mathematical Programming, 25, 1983, 109–121 [71] B De Finetti, Sulle stratificazioni convesse, Annali di Matematica Pura ed Applicata, 30, 1949, 173–183 104 [72] H Tuy, Sules ines galites lines aires, Colloquium Mathematicum, 13, 1964, 107123 [73] B Săollner, Eigenschaften -grobkonvexer Mengen und Funktionen, Diplomarbeit, Universităat Leipzig, 1991 [74] A D Ioffe and V M Tikhomirov, Theory of Extremal Problems, (Tiˆe´ng Nga), Nauka, Moscow, 1974 [75] B N Pshenhishnui, Convex Analysis and Extremal Problems, (Tiˆe´ng Nga), Nauka, Moscow, 1980 [76] M M Vainberg, Variational Method and Method Monotone Operators, (Tiˆe´ng Nga), Nauka, Moscow, 1972 [77] G E Silov, Finite-dimentional Linear Spaces, (Tiˆe´ng Nga), Nauka, Moscow, 1969 ... o.ng d` ung `au Mo˙’ d ¯ˆ `oi `oi, quy hoa.ch to` a h` am lˆ B` to´ an quy hoa.ch lˆ an phu.o.ng v` thˆ o `oi, quy hoa.ch to`an phu.o.ng 1.1 B`ai to´an quy hoa.ch lˆ `oi suy rˆo.ng thˆo... B`ai to´an quy hoa.ch to`an phu.o.ng (P ) • f (x) := Ax, x + b, x → sup, x ∈ D : B`ai to´an quy hoa.ch to`an phu.o.ng (Q) • f (x) := Ax, x + b, x + p(x) → inf, x ∈ D : B`ai to´an quy hoa.ch `oi... d¯`ˆe “B`ai to´an quy hoa.ch lˆ - i.nh l´ - i.nh `oi thˆ `oi, D lˆ o” tr`ınh b`ay D y Kuhn-Tucker cu˙’a b`ai to´an quy hoa.ch lˆ `e d¯iˆ `eu kiˆe.n cu c tri cu˙’a b`ai to´an quy hoa.ch to`an phu.o.ng

Ngày đăng: 24/03/2021, 17:41

TỪ KHÓA LIÊN QUAN

TÀI LIỆU CÙNG NGƯỜI DÙNG

TÀI LIỆU LIÊN QUAN