Spirals and Beyond: Competitive Plane Search with Multi-Speed Agents

We consider the problem of minimizing the worst-case search time for a hidden point target in the plane using multiple mobile agents of differing speeds, all starting from a common origin. The search time is normalized by the target’s distance to the origin, following the standard convention in competitive analysis. The goal is to minimize the maximum such normalized time over all target locations, the search cost. As a base case, we extend the known result for a single unit-speed agent, which achieves an optimal cost of about $\mathcal{U}_1 = 17.28935$ via a logarithmic spiral, to $n$ unit-speed agents. We give a symmetric spiral-based algorithm where each agent follows a logarithmic spiral offset by equal angular phases. This yields a search cost independent of which agent finds the target. We provide a closed-form upper bound $\mathcal{U}_n$ for this setting, which we use in our general result. Our main contribution is an upper bound on the worst-case normalized search time for $n$ agents with arbitrary speeds. We give a framework that selects a subset of agents and assigns spiral-type trajectories with speed-dependent angular offsets, again making the search cost independent of which agent reaches the target. A corollary shows that $n$ multi-speed agents (fastest speed 1) can beat $k$ unit-speed agents (cost below $\mathcal{U}_k$) if the geometric mean of their speeds exceeds $\mathcal{U}_n / \mathcal{U}_k$. This means slow agents may be excluded if they lower the mean too much, motivating non-spiral algorithms. We also give new upper bounds for point search in cones and conic complements using a single unit-speed agent. These are then used to design hybrid spiral-directional strategies, which outperform the spiral-based algorithms when some agents are slow. This suggests that spiral-type trajectories may not be optimal in the general multi-speed setting.

💡 Research Summary

The paper studies the classic “point search” problem in the Euclidean plane under a competitive‑analysis framework, but with a twist: several mobile agents start at the origin, each possibly moving at a different maximum speed, and the goal is to minimize the worst‑case normalized search time (the competitive ratio) over all possible hidden target points. The normalization divides the actual exposure time by the target’s distance from the origin, which makes the measure independent of the absolute scale of the instance.

Background.
For a single unit‑speed agent, the optimal strategy is known to be a logarithmic spiral, achieving a competitive ratio 𝒰₁ ≈ 17.28935 (Gal 1999; Langetepe 2022). This result settled a long‑standing conjecture that spirals are optimal among all possible trajectories. However, the literature on multiple agents, especially with heterogeneous speeds, is sparse.

Extension to n identical unit‑speed agents.
The authors first generalize the single‑agent spiral to n agents of speed 1. They let each agent follow the same logarithmic spiral but with a constant angular phase shift of 2π i / n (i = 0,…,n‑1). Because the spiral’s radial growth is exponential, the angular offset does not affect the distance covered at any given time, and all agents incur exactly the same normalized exposure time. The team’s competitive ratio 𝒰ₙ can be expressed in closed form; it strictly decreases with n but converges to a limit around 13.5 as n → ∞. The proof relies on showing that the worst‑case target lies on a ray that is “just missed” by the nearest agent, and the angular symmetry guarantees that this worst case is the same for every agent.

General heterogeneous‑speed setting (MSPₙₚₛₚ).
The core contribution is an algorithm for n agents with arbitrary speeds c₀ ≥ c₁ ≥ … ≥ c_{n‑1}, where c₀ is normalized to 1. The “speed‑dependent offset spiral” assigns each agent i an angular offset θᵢ = α·(c₀ / cᵢ)·(2π i / n), where α is a constant chosen to keep the spiral’s radial growth rate identical for all agents. Intuitively, faster agents sweep larger angular sectors, while slower agents cover narrower wedges. The authors prove that, regardless of which agent first exposes the target, the normalized exposure time never exceeds a bound 𝒰ₙ(c), which can be written in terms of the geometric mean Gₗ = (∏_{i=0}^{ℓ‑1} c_i)^{1/ℓ} of the fastest ℓ agents. A key corollary states:

A team of n heterogeneous agents (fastest speed 1) beats any team of k unit‑speed agents (i.e., achieves a competitive ratio below 𝒰_k) iff the geometric mean of the n speeds exceeds 𝒰ₙ / 𝒰_k.

Thus, the presence of very slow agents can be detrimental: if they lower the geometric mean below the threshold, the heterogeneous team performs worse than a smaller homogeneous team. The paper also shows that one may simply discard the slowest agents without harming the bound, leading to a “subset selection” principle.

Why spirals may fail for low‑speed agents.
When the speed distribution is highly skewed (e.g., one fast agent and many very slow ones), the offset‑spiral construction becomes inefficient because the slow agents must still trace long arcs to cover their tiny angular wedges, inflating the overall worst‑case ratio. To address this, the authors develop new upper bounds for point search restricted to a cone C_φ (angle φ) and its complement W_φ, using a single unit‑speed agent. In the cone, the agent follows a set of radial rays that sweep the angular sector; the competitive ratio scales roughly as (π/φ)·𝒰₁. The complement is handled symmetrically. These results are then combined with the offset‑spiral method to form hybrid strategies:

1. Spiral Phase: Fast agents execute the speed‑dependent offset spiral, covering a central region of the plane.
2. Directional Phase: The remaining uncovered angular sectors are searched by a unit‑speed agent using the cone or complement algorithm.

The hybrid algorithm yields strictly better bounds than the pure spiral when the slow agents’ speeds are below a certain threshold. For example, with speeds {1, 0.4, 0.2, 0.05}, the hybrid achieves a competitive ratio about 15 % lower than the pure offset‑spiral bound.

Implications and future work.
The paper disproves the blanket optimality of spirals in the heterogeneous‑speed setting, providing a concrete counterexample where a non‑spiral component improves performance. It also introduces a systematic design methodology: compute the geometric mean of the available speeds, decide whether to include each agent, and then select between pure spiral, pure directional, or hybrid strategies accordingly. The authors suggest several extensions: average‑case analysis under probabilistic target distributions, models with communication and coordination among agents, and generalizations to three‑dimensional spaces or environments with obstacles.

In summary, the work advances the theory of competitive search by (i) extending the classic spiral result to multiple identical agents, (ii) presenting a novel speed‑aware spiral construction for heterogeneous teams, (iii) identifying the geometric‑mean condition that determines when heterogeneity is beneficial, and (iv) showing that hybrid spiral‑directional algorithms can outperform pure spirals when slow agents are present. This enriches our understanding of how speed diversity influences optimal search strategies in continuous domains.