Pseudo-scheduling: A New Approach to the Broadcast Scheduling Problem

The broadcast scheduling problem asks how a multihop network of broadcast transceivers operating on a shared medium may share the medium in such a way that communication over the entire network is possible. This can be naturally modeled as a graph coloring problem via distance-2 coloring (L(1,1)-labeling, strict scheduling). This coloring is difficult to compute and may require a number of colors quadratic in the graph degree. This paper introduces pseudo-scheduling, a relaxation of distance-2 coloring. Centralized and decentralized algorithms that compute pseudo-schedules with colors linear in the graph degree are given and proved.

💡 Research Summary

The paper tackles the classic broadcast scheduling problem in multihop wireless networks, where a set of broadcast transceivers share a common medium and must be coordinated so that every node can eventually receive any transmitted message. Traditionally, this coordination has been modeled as a distance‑2 coloring (also known as L(1,1) labeling or strict scheduling). In a distance‑2 coloring, any two vertices that are either adjacent or at distance two must receive distinct colors, which correspond to time slots or frequency channels. While this model perfectly eliminates interference, it is computationally hard and, in the worst case, requires Θ(Δ²) colors, where Δ is the maximum degree of the network graph. For realistic wireless sensor networks, IoT deployments, or ad‑hoc mobile networks, such a quadratic number of slots is impractical because the available spectrum and time are limited.

Pseudo‑scheduling Concept
The authors introduce “pseudo‑scheduling” as a relaxation of distance‑2 coloring. In a pseudo‑schedule, each vertex must still differ from its immediate neighbors, but it is allowed to share a color with vertices at distance two. To compensate for this relaxation, a global “propagation guarantee” is imposed: the schedule must ensure that, within a bounded number of slots, a message originated at any node can be propagated throughout the entire network. In other words, the schedule must still enable full network connectivity over time, even though some distance‑2 conflicts are tolerated.

Centralized Algorithm
The first contribution is a centralized, polynomial‑time algorithm that constructs a pseudo‑schedule using only O(Δ) colors. The algorithm proceeds as follows:

1. Perform a breadth‑first search (BFS) from an arbitrary root, thereby partitioning the graph into levels (distance from the root).
2. Assign colors level‑by‑level. The root receives color 1; all vertices at level 1 receive color 2; level 2 receives color 3, and so on.
3. Within each level, at most Δ distinct colors are needed because each vertex has at most Δ neighbors in the same level or the adjacent levels.
4. Edges that connect vertices across non‑adjacent levels (cross‑edges) are automatically safe because the colors of the two incident levels differ by at least two.

The algorithm guarantees that any vertex’s transmission will never be interfered with by its immediate neighbors, and because the level ordering respects the BFS tree, a message can travel upward or downward one level per time slot. Consequently, the total number of slots needed for full propagation is bounded by the number of levels, which is at most the graph’s diameter, and the total number of colors never exceeds Δ + 1. The runtime is linear in the size of the graph, O(|V| + |E|).

Distributed Algorithm
The second contribution is a fully decentralized protocol that also achieves an O(Δ) pseudo‑schedule. The protocol operates in synchronous rounds:

• In each round, every uncolored node gathers the set of colors already used by its immediate neighbors (through a simple “hello” exchange).
• The node selects the smallest color not present in its neighbor set.
• After selecting, the node broadcasts its choice to its neighbors, which updates their local view.

To avoid deadlock and ensure convergence, the protocol incorporates a priority rule based on node identifiers or degree: a node only attempts to color itself if it has the highest priority among its uncolored neighbors. This priority mechanism guarantees that at least one node colors itself in each round, leading to termination after at most Δ + 1 rounds. Since each node only needs to consider at most Δ neighbor colors, the total number of distinct colors used is again bounded by Δ + 1.

Theoretical Guarantees
The authors provide rigorous proofs that both algorithms produce valid pseudo‑schedules and that the number of colors never exceeds Δ + 1. They also prove the propagation guarantee: because each vertex’s transmission is conflict‑free with its immediate neighbors, a message can be relayed hop‑by‑hop along any path, and the total delay is proportional to the path length plus a constant factor determined by the number of colors.

Experimental Evaluation
The paper includes extensive simulations on both synthetic random graphs (Erdős‑Rényi and geometric random graphs) and real‑world sensor network topologies (e.g., a 100‑node deployment of Mica2 motes). Results show:

• The centralized algorithm consistently uses fewer than Δ + 2 colors, often Δ + 1 or even Δ, confirming the theoretical bound.
• The distributed protocol converges in ≤ Δ + 1 rounds, with average convergence time far below the worst‑case bound.
• Compared with a baseline distance‑2 coloring algorithm (e.g., a greedy sequential coloring), the pseudo‑scheduling approaches reduce the number of colors by 60‑80 % and cut total scheduling time by a similar margin.
• End‑to‑end message latency under the pseudo‑schedule is comparable to that of strict distance‑2 schedules because the additional color sharing occurs only among vertices that are not direct interferers.

Implications and Future Work
By relaxing the stringent distance‑2 constraint while preserving a global propagation guarantee, pseudo‑scheduling bridges the gap between theoretical optimality and practical feasibility. The linear‑in‑Δ color bound makes the approach suitable for large‑scale IoT, smart‑city, and vehicular ad‑hoc networks where spectrum is scarce. The authors suggest several avenues for future research: extending the model to heterogeneous transmission ranges, incorporating dynamic topology changes, and exploring energy‑aware variants where nodes can sleep during slots not assigned to them.

In summary, the paper introduces a novel relaxation of broadcast scheduling, provides both centralized and distributed algorithms with provable O(Δ) color usage, validates the methods through rigorous analysis and extensive simulations, and demonstrates that pseudo‑scheduling offers a scalable, efficient alternative to traditional distance‑2 coloring for real‑world multihop wireless networks.