点上方链接下载,找需要的资源链接
多智能体系统——竞争网络下异构多智能体系统的分组一致性问题 (附论文链接+源码Matlab)
Yu F, Ji L, Yang S. Group consensus for a class of heterogeneous multi-agent networks in the competition systems[J]. Neurocomputing, 2020, 416: 165-171.
研究内容:
研究了连续时间的异构多智能体系统在竞争网络下的分组一致性问题。系统由两类智能体组成——一阶智能体和二阶智能体。与现有研究不同的是,提出了一种新颖的分布式分组一致性控制协议,智能体之间互为竞争的关系。基于稳定性理论、频域方法,讨论了在无时间延迟和有时间延迟影响下的两种系统分组一致性的情形。此外,从理论上建立了实现分组一致性的可容忍的时间延迟上界。通过一系列的数值模拟验证了理论的研究。
系统模型:
{
ξ
˙
i
(
t
)
=
ζ
i
(
t
)
,
ξ
˙
i
(
t
)
=
u
i
(
t
)
,
i
∈
s
^
1
,
ξ
˙
i
(
t
)
=
u
i
(
t
)
,
i
∈
s
^
2
.
\begin{gathered} \begin{cases}\dot{\xi}_{i}(t)=\zeta_{i}(t),\\ \dot{\xi}_{i}(t)=u_{i}(t),\quad i\in\hat{s}_{1},\end{cases} \\ \dot{\xi}_{i}(t)=u_{i}(t),\quad\;\; i\in\hat{s}_{2}. \end{gathered}
{ξ˙i(t)=ζi(t),ξ˙i(t)=ui(t),i∈s^1,ξ˙i(t)=ui(t),i∈s^2.
控制协议(无时间延迟):
对于一阶智能体:
{
ζ
˙
i
(
t
)
=
−
β
[
∑
j
∈
N
i
a
i
j
[
ζ
i
(
t
)
+
ζ
j
(
t
)
]
]
+
ω
i
(
t
)
,
ω
˙
i
(
t
)
=
−
α
[
∑
j
∈
N
i
a
i
j
[
ζ
i
(
t
)
+
ζ
j
(
t
)
]
]
,
∀
i
∈
s
^
2
\begin{cases}{\dot{\zeta}_{i}(t)=-\beta\left[\sum_{j\in N_{i}}a_{i j}[\zeta_{i}(t)+\zeta_{j}(t)]\right]+\omega_{i}(t),}\\ {\dot{\omega}_{i}(t)=-\alpha\left[\sum_{j\in N_{i}}a_{i j}[\zeta_{i}(t)+\zeta_{j}(t)]\right],\quad\forall i\in \hat{s}_{2}}\end{cases}
⎩
⎨
⎧ζ˙i(t)=−β[∑j∈Niaij[ζi(t)+ζj(t)]]+ωi(t),ω˙i(t)=−α[∑j∈Niaij[ζi(t)+ζj(t)]],∀i∈s^2
对于二阶智能体:
{
ξ
i
(
t
)
=
ζ
i
(
t
)
,
ζ
˙
i
(
t
)
=
−
α
[
∑
j
∈
N
i
a
i
j
[
ζ
i
(
t
)
+
ζ
j
(
t
)
]
]
−
β
[
∑
j
∈
N
i
a
j
[
ζ
i
(
t
)
+
ζ
j
(
t
)
]
]
,
∀
i
∈
s
^
1
,
\begin{aligned} &\left\{\begin{array}{l}{{\xi_{i}(t)=\zeta_{i}(t),}}\\ {{\dot{\zeta}_{i}(t)=-\alpha\left[\sum_{j\in\mathcal{N}_{i}}a_{i j}[\zeta_{i}(t)+\zeta_{j}(t)]\right]-\beta\left[\sum_{j\in\mathcal{N}_{i}}a_{j}[\zeta_{i}(t)+\zeta_{j}(t)]\right],}} {{\forall i\in\hat{s}_{1},}}\end{array}\right. \end{aligned}
{ξi(t)=ζi(t),ζ˙i(t)=−α[∑j∈Niaij[ζi(t)+ζj(t)]]−β[∑j∈Niaj[ζi(t)+ζj(t)]],∀i∈s^1,
控制协议(无时间延迟):
对于一阶智能体:
{
ζ
˙
i
(
t
)
=
−
β
[
∑
j
=
N
a
i
j
[
ζ
i
(
t
−
τ
)
+
ζ
j
(
t
−
τ
)
]
]
+
ω
i
(
t
)
,
ω
˙
i
(
t
)
=
−
α
[
∑
j
∈
N
a
j
[
ζ
i
(
t
−
τ
)
+
ζ
j
(
t
−
τ
)
]
]
.
∀
i
∈
s
^
2
.
\begin{cases}\dot{\zeta}_i(t)=-\beta\left[\sum_{j=N}a_{ij}[\zeta_i(t-\tau)+\zeta_j(t-\tau)]\right]+\omega_i(t),\\ \\ \dot{\omega}_i(t)=-\alpha\left[\sum_{j\in\mathcal{N}}a_j[\zeta_i(t-\tau)+\zeta_j(t-\tau)]\right].\quad\forall i\in\hat{s}_2.\end{cases}
⎩
⎨
⎧ζ˙i(t)=−β[∑j=Naij[ζi(t−τ)+ζj(t−τ)]]+ωi(t),ω˙i(t)=−α[∑j∈Naj[ζi(t−τ)+ζj(t−τ)]].∀i∈s^2.
对于二阶智能体:
{
ξ
i
(
t
)
=
ζ
i
(
t
)
,
ζ
˙
i
(
t
)
=
−
α
[
∑
j
∈
N
i
a
i
j
[
ζ
i
(
t
−
τ
)
+
ζ
j
(
t
−
τ
)
]
]
−
β
[
∑
j
∈
N
i
a
j
[
ζ
i
(
t
−
τ
)
+
ζ
j
(
t
−
τ
)
]
]
,
∀
i
∈
s
^
1
,
\begin{aligned} &\left\{\begin{array}{l}{{\xi_{i}(t)=\zeta_{i}(t),}}\\ {{\dot{\zeta}_{i}(t)=-\alpha\left[\sum_{j\in\mathcal{N}_{i}}a_{i j}[\zeta_{i}(t-\tau)+\zeta_{j}(t-\tau)]\right]-\beta\left[\sum_{j\in\mathcal{N}_{i}}a_{j}[\zeta_{i}(t-\tau)+\zeta_{j}(t-\tau)]\right],}} {{\forall i\in\hat{s}_{1},}}\end{array}\right. \end{aligned}
{ξi(t)=ζi(t),ζ˙i(t)=−α[∑j∈Niaij[ζi(t−τ)+ζj(t−τ)]]−β[∑j∈Niaj[ζi(t−τ)+ζj(t−τ)]],∀i∈s^1,
定理:
当满足 α 2 + β 2 ω i 0 2 < ω i 0 2 Re 2 ( λ i ) + Im 2 ( λ i ) , i ∈ s \sqrt{\alpha^2+\beta^2\omega_{i0}^2}<\frac{\omega_{i0}^2}{\sqrt{\text{Re}^2(\lambda_i)+\text{Im}^2(\lambda_i)}},i\in s α2+β2ωi02<Re2(λi)+Im2(λi)ωi02,i∈s, ω i 0 \omega_{i0} ωi0满足 α tan ω i 0 τ = β ω i 0 \alpha\tan\omega_{i0}\tau=\beta\omega_{i0} αtanωi0τ=βωi0,异构多智能体系统可实现渐近时间分组一致性。
时间延迟满足: τ < min λ i ≠ 0 { arctan ( β α ϕ ~ ) ϕ ~ } , i ∈ s . \tau<\min\limits_{\lambda_i\neq0}\bigg\{\dfrac{\arctan(\frac{\beta}{\alpha}\tilde{\phi})}{\tilde{\phi}}\bigg\},i\in s. τ<λi=0min{ϕ~arctan(αβϕ~)},i∈s.
其中 ϕ ~ = ∣ λ i ∣ 2 β 2 + ∣ λ i ∣ 4 β 4 + 4 ∣ λ i ∣ 2 α 2 2 \tilde{\phi}=\sqrt{\frac{|\lambda_i|^2\beta^2+\sqrt{|\lambda_i|^4\beta^4+4|\lambda_i|^2\alpha^2}}{2}} ϕ~=2∣λi∣2β2+∣λi∣4β4+4∣λi∣2α2
//仿真案例
链接🔗:Matlab
进入主页,可以私信