Abstract:This paper offers a data-driven divide-and-conquer strategy to analyze large-scale interconnected networks, characterized by both unknown mathematical models and interconnection topologies. Our data-driven scheme treats an unknown network as an interconnection of individual agents (a.k.a. subsystems) and aims at constructing their symbolic models, referred to as discrete-domain representations of unknown agents, by collecting data from their trajectories. The primary objective is to synthesize a control strategy that guarantees desired behaviors over an unknown network by employing local controllers, derived from symbolic models of individual agents. To achieve this, we leverage the concept of alternating sub-bisimulation function (ASBF) to capture the closeness between state trajectories of each unknown agent and its data-driven symbolic model. Under a newly developed data-driven compositional condition, we then establish an alternating bisimulation function (ABF) between an unknown network and its symbolic model, based on ASBFs of individual agents, while providing correctness guarantees. Despite the sample complexity in existing work being exponential with respect to the network size, we demonstrate that our divide-and-conquer strategy significantly reduces it to a linear scale with respect to the number of agents. We also showcase that our data-driven compositional condition does not necessitate the traditional small-gain condition, which demands precise knowledge of the interconnection topology for its fulfillment. We apply our data-driven findings to two benchmarks comprising unknown networks with an arbitrary, a-priori undefined number of agents and unknown interconnection topologies.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is: How to construct a data - driven divide - and - conquer strategy to analyze large - scale interconnected networks and synthesize control strategies when both the mathematical model of the interconnected network and the interconnection topology are unknown. Specifically, the paper aims to ensure the desired behavior of the unknown network through local controllers and provide formal correctness guarantees.
### Description of the Main Problem
1. **Modeling and Control of Large - Scale Interconnected Networks**:
- The number of subsystems in the interconnected network may be very large, and the dynamic characteristics of these subsystems and their interconnection topologies may be unknown.
- This complexity makes it difficult for traditional methods to be directly applied, especially when an accurate mathematical model is required.
2. **Construction of Symbolic Models**:
- A symbolic model (or finite abstraction) is a discrete representation of the original continuous system and is used to design controllers that meet complex specifications.
- For large - scale interconnected networks, constructing symbolic models faces the problems of high computational complexity and difficult model identification.
3. **Elimination of the Small - Gain Condition**:
- The traditional small - gain condition requires accurate knowledge of the interconnection topology, which is often infeasible in practical applications.
- The paper proposes a data - driven method without the small - gain condition, thereby reducing the dependence on the interconnection topology.
### Solutions
The paper proposes a data - driven divide - and - conquer strategy. The specific steps are as follows:
- **Construction of Symbolic Models at the Subsystem Level**:
- Consider the unknown network as the interconnection of multiple subsystems and construct its symbolic model by collecting the trajectory data of the subsystems.
- Use the Alternating Sub - Bisimulation Function (ASBF) to measure the closeness between the state trajectory of each unknown subsystem and its symbolic model.
- **Data - Driven Composition Conditions**:
- Propose a new data - driven composition condition for constructing the Alternating Bisimulation Function (ABF) of the entire network without relying on traditional interconnection topology knowledge.
- This method significantly reduces the sample complexity, from exponential level to linear level.
- **Control Strategy Synthesis**:
- Synthesize local controllers based on symbolic models and ensure that these local controllers can work together to achieve control of the entire unknown network.
### Formula Representation
- **Alternating Sub - Bisimulation Function (ASBF)**:
\[
V_i(x_i, \hat{x}_i) = rX \sum_{j = 1} q_j^i p_j^i(x_i, \hat{x}_i)
\]
where \( p_j^i(x_i, \hat{x}_i) \) is a user - defined basis function and \( q_j^i \) is an unknown variable.
- **Alternating Bisimulation Function (ABF)**:
\[
V(q, x, \hat{x}) := \sum_{i = 1}^M V_i^*(q_i, x_i, \hat{x}_i)
\]
where \( V_i^* \) is the optimal ASBF for each subsystem.
- **Condition (5.1)**:
\[
\sum_{i = 1}^M (\mu_i^N+\varpi_i^N + L_i\sigma_i) \leq 0
\]
where \( L_i=\max \{L_1^i, L_2^i\} \), and \( L_1^i \) and \( L_2^i \) are Lipschitz constants.
Through these methods, the paper successfully solves the control problem of large - scale interconnected networks and provides formal correctness guarantees.