Abstract:The problem of selecting an optimal seed set to maximise influence in networks has been a subject of intense research in recent years. However, despite numerous works addressing this area, it remains a topic that requires further elaboration. Most often, it is considered within the scope of classically defined graphs with a spreading model in the form of Independent Cascades. In this work, we focus on the problem of budget-constrained influence maximisation in multilayer networks using a Linear Threshold Model. Both the graph model and the spreading process we employ are less prevalent in the literature, even though their application allows for a more precise representation of the opinion dynamics in social networks. This paper aims to answer which of the sixteen evaluated seed selection methods is the most effective and how similar they are. Additionally, we focus our analysis on the impact of spreading model parameters, network characteristics, a budget, and the seed selection methods on the diffusion effectiveness in multilayer networks. Our contribution also includes extending several centrality measures and heuristics to the case of such graphs. The results indicate that all the factors mentioned above collectively contribute to the effectiveness of influence maximisation. Moreover, there is no seed selection method which always provides the best results. However, the seeds chosen with VoteRank-based methods (especially with the $v-rnk-m$ variant we propose) usually provide the most extensive diffusion.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the optimal seed selection method for influence maximization using the Linear Threshold Model (LTM) in multi - layer networks under budget constraints. Specifically, the paper focuses on the following points:
1. **Effectiveness of seed selection methods**: The paper evaluates 16 different seed selection methods to determine which one is the most effective in multi - layer networks. These methods include centrality - based metrics (such as Degree Centrality, Betweenness Centrality, Closeness Centrality, etc.) and other heuristic methods (such as VoteRank, etc.).
2. **Influence of propagation model parameters**: The influence of different propagation model parameters on the diffusion effect is studied. For example, the influence of the threshold \(\mu\) in LTM and the parameter \(\delta\) in the protocol function on the diffusion process.
3. **Influence of network characteristics and budget**: The influence of network characteristics (such as the size, number of layers and topological structure of the network) and budget constraints on seed selection and diffusion effect is analyzed.
4. **Similarities of seed selection methods**: The similarities and differences between different seed selection methods and their performance under different conditions are explored.
5. **Evaluation of diffusion effect**: The comprehensive influence of various factors (such as network characteristics, budget, seed selection methods, etc.) on the final diffusion effect is experimentally evaluated.
The main contribution of the paper is to extend several centrality metrics and heuristic methods to make them applicable to multi - layer networks, and experimentally prove the effectiveness of these methods in multi - layer networks. In particular, the VoteRank - based method (especially the proposed v - rnk - m variant) can generally achieve the widest diffusion effect.
### Formula Explanation
- **Linear Threshold Model (LTM)**:
- Each node \(v\) selects a threshold \(\mu_v\), randomly selected from the uniform distribution \(U(0, 1)\).
- The condition for node \(v\) to become active at time step \(t\) is that the sum of the influence of its neighbors at time step \(t - 1\) exceeds \(\mu_v\):
\[
\sum_{u\in U\cap S_{t - 1}}w_{u, v}\geq\mu_v
\]
- **Multilayer Linear Threshold Model (Multilayer LTM)**:
- Each node \(v\) can only be influenced by its neighbors \(U\), and the connection weight \(w_{u, v}=\frac{1}{\text{degree}(v)}\), so that \(\sum_{u\in U}w_{u, v} = 1\).
- The input \(y_v(t)\) received by node \(v\) at time step \(t\) is:
\[
y_v(t)=\begin{cases}
1, & \text{if }\sum_{u\in U\cap S_{t - 1}}w_{u, v}>\mu\\
0, & \text{otherwise}
\end{cases}
\]
- The average input \(y_a(t)\) of actor \(a\) at time step \(t\) is:
\[
y_a(t)=\frac{1}{|K|}\sum_{k\in K}y_{v_k}(t)
\]
- The state \(x_a(t)\) of actor \(a\) at time step \(t\) is determined by the protocol function:
\[
x_a(t)=\begin{cases}
1, & \text{if }y_a(t)\geq\delta\text{ or }x_a(t - 1)=1\\
0, & \text{otherwise}
\end{cases}
\]