Information Coverage Rate (ICR)

• Information Coverage Rate (ICR) is a metric that quantifies information spread by measuring both active nodes and informed-but-inactive participants in network systems.
• ICR employs domain-specific models such as IC/LT diffusion in social networks, throwbox methodologies in DTNs, and product models in wireless communications for performance optimization.
• ICR’s submodular properties and algorithmic strategies like greedy and heuristic methods facilitate efficient estimation and comparative analysis in complex stochastic environments.

Information Coverage Rate (ICR) quantifies the extent to which information permeates a networked system, capturing not only directly touched or "active" entities but also a broader set of informed or functionally covered participants. The ICR concept arises independently in social influence diffusion, opportunistic/mobile networking, and wireless communication, each with domain-specific mathematical and algorithmic frameworks. Across all domains, ICR rigorously measures the joint effect of information reach (coverage) and a secondary objective such as activation, reliability, or rate, thereby facilitating optimization and comparative analysis in complex stochastic environments.

1. Formal Definitions and Model-Specific Instantiations

ICR's precise definition is domain-dependent:

• Social Networks (IC/LT diffusion): For a social graph $G=(V,E)$ of size $n=|V|$, the Information Coverage Function for a seed set $S\subseteq V$ under the Independent Cascade (IC) or Linear Threshold (LT) diffusion model is

$F(S) = \mathbb{E}[|A|] + \mathbb{E}[|L|]$

where $A$ is the set of ultimately active nodes and $L$ is the set of informed-but-inactive nodes (those having at least one active in-neighbor at any time during the process). The ICR is then

$$\mathrm{ICR}(S) = \frac{F(S)}{n} = \frac{\mathbb{E}[|A|] + \mathbb{E}[|L|]}{n}$$

Optionally, a weighted variant $W(S) = \mathbb{E}[|A|] + \lambda \mathbb{E}[|L|]$ interpolates between classical influence (λ=0) and full information coverage (λ=1) (Wang et al., 2015).

• Throwbox-augmented Delay-Tolerant Networks (DTNs): Information Coverage Rate is the stabilized asymptotic fraction of distinct places (throwboxes) ever hosting a message; $G_d(p)/N$, with $G_d(p)$ the number of covered places and $N=|P|$ the total number of places (Saha et al., 2016).
• Wireless Networks: In coexisting THz/RF femto- and macro-cell architectures, ICR is defined as the product of coverage probability and average rate:

$\mathrm{ICR} = P_c \times \bar{R}$

where $P_c$ is the probability that $\mathrm{SINR} > \theta$ at a typical user, and $\bar{R}$ is the ergodic per-user throughput (in nats/sec or bits/sec) (Kouzayha et al., 2021). In sequential transmissions with reliability constraints, ICR is the maximum reliable throughput subject to $n$-message coverage probability constraints, i.e., $R^*(\varepsilon, n)$ (Park et al., 2017).

2. Theoretical Properties and Submodularity

Rigorous analysis of ICR reveals underlying mathematical structure:

• Submodularity in Social Influence: $F(S)$ is submodular in $S$ under both IC and LT models. For any live-arc realization $G_L$,

$C_{G_L}(S) = |R_{G_L}(S)| + \lambda |U_{G_L}(S)|$

is submodular; thus, so is $F(S)$ and weighted $W(S)$, permitting (1-$1/e$)-approximate greedy maximization (Wang et al., 2015).

• Phase Transition and Component Size in BNW: In DTN bipartite models, the main result is that, asymptotically, all degree-$\ge$1 places belong to a single giant component whose size gives ICR. Explicit formulae relate ICR to cumulative degree distributions and key network parameters (Saha et al., 2016).
• Integral and Optimization Structure in Wireless: ICR functionals are quasi-concave in deployment parameters such as THz AP fraction and bias. First-order conditions characterize ICR-optima, requiring joint (gradient-based) optimization (Kouzayha et al., 2021).

3. Algorithmic Computation and Optimization

Domain-specific algorithms address the NP-hardness or combinatorial complexity of maximizing ICR:

• Greedy + Lazy-Update (Social Influence): Achieves $(1-1/e - \epsilon)$-approximation to $\max_{|S|=k} W(S)$. Marginal gains computed via Monte Carlo, lazy-recomputing $\Delta(v)$ amortizes computation (Wang et al., 2015).
• Effective-Degree Heuristic: Prioritizes nodes with high (discounted) out-degree. Fast $O(k(n+m))$ but heuristically suboptimal relative to greedy (Wang et al., 2015).
• Bipartite Network Model (DTN): Analytical expressions for ICR allow for closed-form computation; the giant component size $G_b$ reads

$$G_b = N \Big[1 - \left(1 - \frac{v}{\mu(\mu-1)}\right)^{N-1}\Big]$$

after mapping protocol parameters to model variables. Empirical validation confirms accuracy (Saha et al., 2016).

• Numerical Optimization (Wireless): Integral expressions for $P_c$ and $\bar{R}$ require evaluation under system parameters $(\delta_T, B_T)$, with global maxima for ICR found via root-finding, e.g., Newton–Raphson (Kouzayha et al., 2021).
• Reliability-Constrained Rate Optimization: In sequence-reliable downlink, closed-form bounds for $R^*(\varepsilon, n)$ can be obtained by inverting tight power-law outer bounds for the $n$-successive SIR coverage probability (Park et al., 2017).

4. Empirical Findings and Parameter Sensitivity

Experimental and analysis-based findings illuminate the behavior and limiting regimes of ICR:

• In social networks, informed nodes may outnumber active ones by 30–50% for small seed budgets; ICR saturates with increased seeding, confirming diminishing returns (Wang et al., 2015).
• In throwbox DTNs, ICR falls off as $1/N^2$–$1/N^3$ with increasing place count; coverage increases quadratically with agent activity parameter $\mu$ and linearly decreases with buffer refresh probability $p$. Super-preferential mobility (high clustering exponent $\alpha$) degrades ICR severely (Saha et al., 2016).
• In coexisting THz/RF wireless, increasing $\delta_T$ (fraction THz APs) increases rate but decreases coverage, manifesting a clear tradeoff. Edge users require lower $\delta_T$ and bias for robust coverage, while central users can exploit high $\delta_T$ for maximum throughput (Kouzayha et al., 2021).
• In downlink-sequence wireless, reliability-constrained ICR tightens as reliability target $\eta \to 1$, with closed-form approximations becoming exact in this regime (Park et al., 2017).

5. Practical Estimation and Implementation Considerations

Applying ICR in real-world systems requires:

• Social Diffusion: Estimating $F(S)$ via repeated diffusion simulations (Monte Carlo), tracking the partition of active and informed sets per realization. For high precision, on the order of thousands of runs is typical; algorithmic seed selection uses the same simulations for marginal gain estimation (Wang et al., 2015).
• DTN/Throwbox Systems: Coverage can be computed directly from model parameters using closed-form equations. Empirical traces processed into visit sequences further validate the model; parameter calibration is performed using synthetic or real mobility traces (Saha et al., 2016).
• Wireless Networks: Both $P_c$ and $\bar{R}$ require either closed-form integration (where possible) or numerical techniques for finite domains and multi-tier settings. ICR maximization is inherently joint in system parameters—fraction of high-rate nodes, biasing parameters, etc.—and dependent on user location statistics (Kouzayha et al., 2021). In reliability-constrained downlink, invertible expressions for ICR (reliable throughput) enable rapid calculation for adaptation and control (Park et al., 2017).

6. Comparative Table of ICR Formalizations

Domain/Model	ICR Expression	Optimization Approach
Social diffusion (IC/LT)	$\frac{\mathbb{E}[|A|] + \mathbb{E}[|L|]}{n}$	Greedy+Lazy, Degree Heuristic
DTN Throwbox (BNW)	$G_d(p)/N$, with $G_d(p)$ from closed-form	Direct computation via analytic formula
THz/RF Wireless	$P_c \times \bar{R}$	Numerical root-finding / integration
Sequence coverage (downlink)	$R^*(\varepsilon, n)$ under $n$-msg reliability, in closed-form approximation	Analytical inversion of coverage bounds

7. Conceptual and Methodological Significance

ICR extends classical notions such as influence spread and coverage probability by integrating qualitatively distinct aspects—awareness, functional contact, joint rate–reliability, or geographic extent—into a single, rigorous performance criterion. This allows for unified optimization frameworks that account for both propagation reach and secondary operational constraints (rate, reliability, buffer refresh, etc.), facilitating both theoretical analysis and practical deployment strategy design.