Dynamic Spectrum Sharing in Next-Gen Wireless Networks

• Dynamic Spectrum Sharing is a method that enables real-time, context-aware reallocation of contiguous spectrum blocks to maximize wireless network efficiency amid variable demand and interference.
• It employs algorithmic approaches like graph coloring and bandwidth–coverage sorting to strategically allocate spectrum while balancing feasibility and transmitter deployment.
• This technique is essential for next-generation networks such as 6G, supporting heterogeneous deployments and paving the way for adaptive, machine learning-driven enhancements.

Dynamic Spectrum Sharing (DSS) is a paradigm enabling real-time, context-aware reallocation of spectrum among heterogeneous wireless systems, with the goal of exploiting spectral opportunities and increasing utilization efficiency under time-varying demand and interference constraints. In next-generation wireless networks, including 6G, DSS is seen as essential for addressing the proliferation of dense, diverse, and high-throughput applications, accommodating both primary (licensed) and secondary (opportunistic or lightly licensed) users, and adapting to spatial, temporal, and frequency heterogeneity in network conditions (Walishetti et al., 2024).

1. Fundamental Concepts and System Models

DSS departs from static, long-term spectrum assignments, instead allowing spectrum access rights to dynamically adjust in response to operational context, demand, and policy constraints. A core task is the allocation of limited, contiguous frequency bands to wireless base stations (BSs) or users so as to maximize the number of active transmitters, total coverage, and aggregate bandwidth utilization, subject to strict interference avoidance between overlapping coverage areas. This is often formalized via an interference (or conflict) graph $G=(V,E)$, where each vertex represents a BS, and undirected edges link BSs whose coverage regions overlap—thereby necessitating assignment of non-overlapping spectral resources (Walishetti et al., 2024).

Each DSS allocation must respect:

• Contiguity constraint: Every BS is assigned a contiguous block of $B_i$ units in a total pool of $F$ units: $\{f_i, f_i+1, ..., f_i+B_i-1\}$, with $f_i \in [1, F-B_i+1]$.
• Non-interference (graph coloring) constraint: For every $(i,j)\in E$, the BSs’ allocations must be disjoint: $$\{f_i,...,f_i+B_i-1\} \cap \{f_j,...,f_j+B_j-1\} = \emptyset$$ (Walishetti et al., 2024).

Performance is measured using several standardized metrics:

• Feasibility Indicator (FI): $FI=1$ if all BSs' allocations fit within the available spectrum, $FI=0$ otherwise.
• Bandwidth Usage (BU): The highest index of any assigned band.
• Coverage Area (CA): Aggregate area served, summing over allocated BSs, corrected for edge clipping.
• Bandwidth–Coverage Product (BC): $\sum_{i} R_i B_i$ for all allocated BSs, combining spatial and spectral usage.
• Total Transmitters while Feasible (TF): Number of BSs allocated before spectrum exhaustion.

DSS methodologies aim to maximize combinations of these metrics based on application objectives and spectrum constraints (Walishetti et al., 2024).

2. Algorithmic Approaches to DSS

DSS employs algorithmic ordering of BSs or users to optimize allocation under interference constraints. Five representative DSA (Dynamic Spectrum Access) algorithms include:

1. Most-Overlaps Sort (Welsh–Powell Algorithm): Orders BSs by descending degree in the interference graph (highest number of conflicts first), allocating to "hard" BSs earlier—equivalent to greedy graph coloring (Walishetti et al., 2024).
2. Bandwidth–Coverage Sort: Orders by descending product $R_i B_i$ (coverage × bandwidth), prioritizing maximum BC.
3. Least-Bandwidth Sort: Orders by ascending $B_i$, packing small-bandwidth BSs first to maximize TF.
4. Least-Coverage Sort: Orders by ascending $R_i$, similar effect to least-bandwidth, targeting more allocations.
5. Random Sort: Allocates in uniform random order.

All algorithms sort BSs and allocate in $O(N\log N)$ time, then sequentially assign the lowest-possible contiguous spectrum block to each BS, ensuring no violations of the coloring constraint (Walishetti et al., 2024).

Simulation evidence shows that:

• Feasibility-prioritized sorts (Most-Overlaps, Bandwidth–Coverage) maximize $FI$ and BU efficiency when spectrum is abundant but pack fewer total transmitters and less coverage under tight constraints.
• Least-Bandwidth and Least-Coverage sorts maximize total transmitters and coverage area when spectrum is constrained, but seldom achieve $FI=1$.
• Trade-offs emerge between all-or-nothing (feasibility-optimized) and pack-small-first (utilization-optimized) strategies; random order serves as a baseline (Walishetti et al., 2024).

3. Evaluative Metrics and Trade-offs

DSS must be evaluated using multiple metrics that reflect both spectrum utilization efficiency and service coverage:

Metric	Formal Definition	Significance
FI	$FI=1$ if $\forall i,\ f_i+B_i-1 \leq F$; $0$ else	Feasibility of allocation
BU	$\max_i (f_i+B_i-1)$	Used bandwidth
CA	$\sum_{i\in\mathcal{N}} \pi R_i^2 C_i$	Total coverage area
BC	$\sum_{i\in\mathcal{N}} R_i B_i$	Combined coverage-bandwidth
TF	Max $n$ with $f_i+B_i-1 \leq F$	Total BSs that can be served

Performance analysis across variable $N$ (number of BSs), $F$ (bandwidth), and distributions of $B_i$, $R_i$ demonstrates that no single sorting rule is optimal across all metrics. In spectrum-scarce regimes, maximizing the number of allocated transmitters and spatial coverage typically conflicts with maximizing likelihood of full-network feasibility (Walishetti et al., 2024).

4. Practical Applications and Extensions

The presented DSS framework is immediately relevant for spectrum management in 6G networks, where dense deployments of heterogeneous wireless devices (with varying coverage radii and bandwidth demands) are expected across both licensed and shared bands. Applications include cellular base station deployments, ultra-dense small cell networks, and industrial/private LTE/5G/6G scenarios.

Immediate extensions include:

• Non-contiguous spectrum allocation: Allowing fragmented assignments to enhance packing efficiency.
• Directional antennas and mobility: Incorporating spatial selectivity and time-varying coverage areas.
• Machine learning and online adaptation: Employing RL or learning-based sorting priorities, responsive to ongoing changes in network state and traffic demand.
• Fairness and service guarantees: Balancing allocation to maximize both spectral efficiency and user/service diversity (Walishetti et al., 2024).

5. Recommendations and Design Insights

Insights from simulation demonstrate that:

• In spectrum-rich conditions, algorithms prioritizing feasibility (e.g., Most-Overlaps, Bandwidth–Coverage) ensure comprehensive service but potentially underutilize spectrally efficient opportunities in highly constrained environments.
• In spectrum-limited scenarios, least-demand first algorithms (Least-Bandwidth, Least-Coverage) substantially increase the number of successfully allocated transmitters and the total covered area, at the expense of full-network feasibility.
• To balance objectives, hybrid ordering (e.g., demand-tiered allocation) or adaptive real-time selection of sorting rules provides a path forward for future DSS systems.

These results inform both the design of next-generation spectrum management protocols and regulatory policy, emphasizing the necessity of dynamic, context-aware, and adaptive allocation mechanisms (Walishetti et al., 2024).

6. Research Directions and Open Challenges

Several directions for advancing DSS are identified:

• Enabling non-contiguous spectrum allocations and dynamic fragmentation, particularly relevant for fragmented mid-band and mmWave deployments.
• Designing algorithms robust to variable and directional coverage areas, including in ultra-dense urban or industrial environments.
• Integrating ML and RL approaches to dynamically adapt allocation strategies to observed interference, traffic, and spectrum availability.
• Addressing time-varying demand and real-time enforcement of interference constraints.
• Developing scalable, distributed decision-making procedures, especially for large and heterogeneous network deployments (Walishetti et al., 2024).

Rich opportunities exist for interdisciplinary advances at the intersection of graph theory, optimization, and online learning to further increase DSS efficiency and robustness.