线性化Vicsek模型的同步性分析.pdfVIP

Proceedings of the 25th Chinese Control Conference 7–11 August, 2006, Harbin, Heilongjiang Vicsek , , E-mail: Lguo@ 100080 : Vicsek : , , Vicsek , , Laplace , Convergence Analysis of Linearized Vicsek’s Model Gongguo Tang, Lei Guo Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100080 E-mail: Lguo@ Abstract: In this paper, the linearized Vicsek’s Model will be considered. A suf?cient condition will be established to guarantee the convergence of the agents’ ?y headings in large probability when the number of agents in the model is suf?ciently large. Key Words: multi-agent system , collective behavior, linearized Vicsek’s Model, convergence in large probability, graph Laplacians , normalized algebraic connectivity 1 (Introduction) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11] [1, 2, 3, 4] T.Vicsek [7] [5, 6] r T.Vicsek (spontaneous symmetry breaking) 1987 C.Reynolds Boid [7] Boid IEEE Catalog Number: 06EX131060334040. , Vicsek : 379 Boid A.Jadbabaie 2003 [8] Vicsek [9] Vicsek vn k ? ? xk(t) = xk(t ? 1) + vn cos θk(t ? 1) ? yk(t) = yk(t ? 1) + vn sin θk(t ? 1) zk(t) = zk(t ? 1) + vneiθk(t?1) Vicsek [8] (1) F.Cucker S.Smale [11] [8, 9, 11] [8] 2 Vicsek ), t?1 k (Problem Formulation) x?y n 1, 2, · · · , n n t (, t Nk(t) = {j : (xk(t) ? xj(t))2 + (yk(t) ? yj(t))2 rn2 } xk(t), yk(t) x?y rn 0 #Nk (t) x?y √ zk(t) = xk(t) + iyk(t), i = ?1 Vicsek θk nk(t) = θk(t) = arctan j ∈Nk (t?1) j ∈Nk (t?1) sin cos θj (t θj (t ? ? 1) 1) . θk (t) = 1 nk(t ? 1) θj(t ? 1) j ∈Nk (t?1) (2) k≤ n} t=0 n} n (?π, π) tn n (1) (2) ) G(n, rn, t) G(t) [12] G k {zk(0), 1 ≤ {θk(0), 1 ≤ k ≤ n Nk (t) ( dk dmin = min 1≤k≤n dk dmax = max 1≤k≤n dk T = diag(dk)1≤k≤n A Laplace L Laplace L = T ?1/2LT ?1/2 P = T ?1A L n 0 = λ0 λ1 · · · λn?1 λ1 λˉ = max1≤k≤n?1 |1 ? λk| 0 ηn 1 k Rk = {z : (1?ηn)rn |z ?zk(0)| (1+ηn)rn} G(n, rn, t) Rk Rmax = max1 k n Rk tn n t t=0 Z(t) = (z1(t), z2(t), ..., zn(t))T , θ(t) = (θ1(t), θ2(t), ..., θn(t))T (1) (2) ? ???? θ(t) = P (t ? 1)θ(t ? 1) ????