马逸文. 多智能体系统滞后一致性研究[D].西南民族大学,2021.DOI:10.27417/d.cnki.gxnmc.2021.000193.
文章目录
- 第一章 绪论
- 第三章 二阶离散线性多智能体系统滞后一致性
- 3.5 数值仿真
第一章 绪论
第三章 二阶离散线性多智能体系统滞后一致性
3.5 数值仿真
x 0 ( k + 1 ) = A x 0 ( k ) + v 0 ( k ) v 0 ( k + 1 ) = B v 0 ( k ) \begin{aligned} x_0(k+1) &= A x_0(k) + v_0(k) \\ v_0(k+1) &= B v_0(k) \end{aligned} x0(k+1)v0(k+1)=Ax0(k)+v0(k)=Bv0(k)
x i ( k + 1 ) = A x i ( k ) + v i ( k ) v i ( k + 1 ) = B v i ( k ) + a ∑ j = 1 n w i j ( x j ( k ) − x i ( k ) ) − b ( x i ( k ) − x 0 ( k − τ ) ) − c ( v i ( k ) − v 0 ( k − τ ) ) \begin{aligned} x_i(k+1) &= A x_i(k) + v_i(k) \\ v_i(k+1) &= B v_i(k) &+ a \sum_{j=1}^{n} w_{ij} (x_j(k)-x_i(k)) \\ &&- b (x_i(k) - x_0(k-\tau)) \\ &&- c (v_i(k) - v_0(k-\tau)) \end{aligned} xi(k+1)vi(k+1)=Axi(k)+vi(k)=Bvi(k)+aj=1∑nwij(xj(k)−xi(k))−b(xi(k)−x0(k−τ))−c(vi(k)−v0(k−τ))
利用基础参数得到的结果如下,对应程序 Main.m
考虑时滞
τ
=
0.15
\tau = 0.15
τ=0.15 的情况,对应程序 Main_Tau.m
在上述基础上,修改
A
=
[
−
0.5
−
0.75
1
−
0.5
]
A = \left[\begin{matrix} -0.5 & -0.75 \\ 1 & -0.5 \\ \end{matrix}\right]
A=[−0.51−0.75−0.5],对应程序 Main_Tau1.m
,结果为
在上述基础上,修改
A
=
[
1
−
1.25
−
1
−
1
]
A = \left[\begin{matrix} 1 & -1.25 \\ -1 & -1 \\ \end{matrix}\right]
A=[1−1−1.25−1],对应程序 Main_Tau2.m
,结果为
在上述基础上,修改
A
=
[
0.5
−
0.25
1
0.5
]
A = \left[\begin{matrix} 0.5 & -0.25 \\ 1 & 0.5 \\ \end{matrix}\right]
A=[0.51−0.250.5],对应程序 Main_Tau3.m
,结果为
在上述基础上,修改
A
=
[
0.5
1.5
1
0.5
]
A = \left[\begin{matrix} 0.5 & 1.5 \\ 1 & 0.5 \\ \end{matrix}\right]
A=[0.511.50.5],对应程序 Main_Tau4.m
,结果为
可以看到结果并不收敛,同时发现特征根并不与论文一致,计算出来的特征根为 λ 1 = 1.7247 , λ 2 = − 0.7247 \lambda_1 = 1.7247, \lambda_2 = -0.7247 λ1=1.7247,λ2=−0.7247。
自己尝试了一下矩阵,将 A A A 改为 A = [ 0.5 0.15 0.1 0.5 ] A = \left[\begin{matrix} 0.5 & 0.15 \\ 0.1 & 0.5 \\ \end{matrix}\right] A=[0.50.10.150.5] 后,效果还算理想,效果如下。
x 0 ( k + 1 ) = A x 0 ( k ) + v 0 ( k ) v 0 ( k + 1 ) = B v 0 ( k ) + r 0 ( k ) r 0 ( k + 1 ) = C r 0 ( k ) \begin{aligned} x_0(k+1) &= A x_0(k) + v_0(k) \\ v_0(k+1) &= B v_0(k) + r_0(k) \\ r_0(k+1) &= C r_0(k) \end{aligned} x0(k+1)v0(k+1)r0(k+1)=Ax0(k)+v0(k)=Bv0(k)+r0(k)=Cr0(k)
x i ( k + 1 ) = A x i ( k ) + v i ( k ) v i ( k + 1 ) = B v i ( k ) + r i ( k ) r i ( k + 1 ) = C r i ( k ) + a ∑ j = 1 n w i j ( x j ( k ) − x i ( k ) + v j ( k ) − v i ( k ) ) − b ( x i ( k ) − x 0 ( k − τ ) ) − c ( v i ( k ) − v 0 ( k − τ ) ) − d ( r i ( k ) − r 0 ( k − τ ) ) \begin{aligned} x_i(k+1) &= A x_i(k) + v_i(k) \\ v_i(k+1) &= B v_i(k) + r_i(k) \\ r_i(k+1) &= C r_i(k) &+ a \sum_{j=1}^{n} w_{ij} (x_j(k)-x_i(k) + v_j(k)-v_i(k)) \\ &&- b (x_i(k) - x_0(k-\tau)) \\ &&- c (v_i(k) - v_0(k-\tau)) \\ &&- d (r_i(k) - r_0(k-\tau)) \end{aligned} xi(k+1)vi(k+1)ri(k+1)=Axi(k)+vi(k)=Bvi(k)+ri(k)=Cri(k)+aj=1∑nwij(xj(k)−xi(k)+vj(k)−vi(k))−b(xi(k)−x0(k−τ))−c(vi(k)−v0(k−τ))−d(ri(k)−r0(k−τ))