Parallel Bayesian Search with No Coordination
read more
Citations
Weighted Group Search on a Line
Multi-round cooperative search games with multiple players
Probabilistically Faulty Searching on a Half-Line
Algorithms for p-Faulty Search on a Half-Line
Probabilistically Faulty Searching on a Half-Line
References
The Theory of Search Games and Rendezvous
Searching in the Plane
Distributed Computing by Oblivious Mobile Robots
On the linear search problem
Related Papers (5)
Frequently Asked Questions (9)
Q2. Why is the term linear search used?
The terminology linear search comes from the fact that the boxes are linearly ordered, and must ideally be checked in that order.
Q3. How is the time to check a key in cryptography?
in cryptography, an attack is better proceeded by systematically checking smaller keys than longer ones, because the time to check a key is typically exponential in its size.
Q4. Why is Acoord a distribution over deterministic search algorithms?
Because the authors allow coordination, any randomized search algorithm is centralized, and thus can be seen as a distribution over deterministic search algorithms.
Q5. what is the probability that a non-coordinating algorithm did not check box bx?
In essence, the authors show that if a non-coordinating algorithm is c-competitive, then it is also c-competitive under disordering of the boxes.
Q6. What is the definition of a functional view of an algorithm?
Definition 3. Given a non-coordinating search algorithm A, the functional view of A is the function N : N+ × N → [0, 1] defined as N(x, t) = Pr[Bx was not yet checked by time t by searcher si] where si is an arbitrary searcher performing A.
Q7. what is the function f(x) = (x)1/(k?
To calculate α, the authors use what the authors know from Eq. (13) about what N looks like, and the authors rely on the refinement of Eq. (11) given by the Presentation Lemma, i.e., that for all t, ∫∞ 1 (1 − N(x, t)) dx = t. Again, fix a t and examine the function f(x) = αρ(x)−1/(k−1).
Q8. What is the probability that none of the k searchers checked x by time t?
the probability that none of the k searchers checked x by time t is N(x, t)k, and thus, by Eq. (6),T(A, x) = ∞∑ t=0 N(x, t)k.
Q9. How does the algorithm perform in a probabilistic setting?
it turns out that the two settings (Bayesian search and linear search) are highly related, as far as order-invariant algorithms are concerned: an order-invariant algorithm that works well against a treasure placed arbitrarily would also work well in any probabilistic setting (under the assumption that the indices of the boxes are ordered according to their relative likelihood).