Best-choice Problem Research Articles

Running minimum in the best-choice problem

The full-information best choice problem asks one to find a strategy maximising the probability of stopping at the minimum (or maximum) of a sequence X_1,cdots ,X_n of i.i.d. random variables with continuous distribution. In this paper we look at more general models, where independent X_j’s may have different distributions, discrete or continuous. A central role in our study is played by the running minimum process, which we first employ to re-visit the classic problem and its limit Poisson counterpart. The approach is further applied to two explicitly solvable models: in the first the distribution of the jth variable is uniform on {j,cdots ,n}, and in the second it is uniform on {1,cdots , n}.

Trapping the Ultimate Success

We introduce a betting game where the gambler aims to guess the last success epoch in a series of inhomogeneous Bernoulli trials paced randomly in time. At a given stage, the gambler may bet on either the event that no further successes occur, or the event that exactly one success is yet to occur, or may choose any proper range of future times (a trap). When a trap is chosen, the gambler wins if the last success epoch is the only one that falls in the trap. The game is closely related to the sequential decision problem of maximising the probability of stopping on the last success. We use this connection to analyse the best-choice problem with random arrivals generated by a Pólya-Lundberg process.

Optimal rules for the sequential selection of uniform spacings

ABSTRACT The unit interval is broken at random into n spacings with the breaking points given by a random sample of size n−1 from the uniform distribution on . A decision-maker observes the lengths of these spacings sequentially and must decide either to select the present spacing or to reject it and continue to observe the next one. No recall of preceding observations is permitted. In this paper, we first find an optimal stopping rule to maximize the probability of selecting the largest spacing. Furthermore, we establish a connection to the classical secretary problem, from which we derive a lower bound for the maximum probability. Moreover, we conjecture that the limiting optimal probability is the same as in the full-information best-choice problem. Next, we find an optimal stopping rule to maximize the expected length of the selected spacing. It is shown that the maximum obtainable expected length lies between the expected value of the second-largest and the largest spacing.

Cost-Efficient Strategy in Clouds with Spot Price Uncertainty

We address a cloud spot bidding problem for user cost optimization. We propose stochastic optimization model to minimize the expected resource rental cost in the presence of spot price uncertainty. The model is based on the well-known full-information best-choice problem. Based on the model, we derive the strategy for cloud spots bidding. The strategy allows to minimize the expected cost for a spot instance in a specific period of time with quality of service guarantee. Our simulation analysis based on realistic settings clearly demonstrates the advantages of the proposed optimization solutions.

On multilateral incomplete information decision models

AbstractThis paper treats the decision problem related to theobservation of a Markov process by decision makers. The informationdelivered to the players is based on the aggregation of thehigh‐frequency data by some functions. Admissible strategies arestopping moments related to the available information. The paymentsare defined by the state at the time of stopping. The players' decision to stop has various effects which depend on the decisionmakers' type. The type β player's stopping decision assignsthe state of the process with chance β, and it offers thisstate to the opponent with probability . The knowledgeabout the type of the players is not public and in this way, thepayers have also different information. The details of thedescription allow to formulate the problem as a Bayesian game withsets of strategies based on the stopping times. It is an extensionof Dynkin's game related to the observation of a Markov process withthe random assignment mechanism of states to the players. Some examples related to the best choice problem (BCP) are analyzed. MSC (2000) Primary: 90D15; Secondary: 93C30.

Iterated full information secretary problem

We consider a full information best-choice problem where an administrator who has only one on-line choice in m consecutive searches has to choose the best candidate in one of them.

Percolation and best-choice problem for powers of paths

AbstractThe vertices of thekth power of a directed path withnvertices are exposed one by one to a selector in some random order. At any time the selector can see the graph induced by the vertices that have already appeared. The selector's aim is to choose online the maximal vertex (i.e. the vertex with no outgoing edges). We give upper and lower bounds for the asymptotic behaviour ofpn,kn1/(k+1), wherepn,kis the probability of success under the optimal algorithm. In order to derive the upper bound, we consider a model in which the selector obtains some extra information about the edges that have already appeared. We give the exact asymptotics of the probability of success under the optimal algorithm in this case. In order to derive the lower bound, we analyse a site percolation process on a sequence of thekth powers of a directed path withnvertices.

Urn sampling distributions giving alternate correspondences between two optimal stopping problems

Abstract The best-choice problem and the duration problem, known as versions of the secretary problem, are concerned with choosing an object from those that appear sequentially. Let (B,p) denote the best-choice problem and (D,p) the duration problem when the total number N of objects is a bounded random variable with prior p=(p1, p2,...,pn) for a known upper bound n. Gnedin (2005) discovered the correspondence relation between these two quite different optimal stopping problems. That is, for any given prior p, there exists another prior q such that (D,p) is equivalent to (B,q). In this paper, motivated by his discovery, we attempt to find the alternate correspondence {p(m),m≥0}, i.e. an infinite sequence of priors such that (D,p(m-1)) is equivalent to (B,p(m)) for all m≥1, starting with p(0)=(0,...,0,1). To be more precise, the duration problem is distinguished into (D1,p) or (D2,p), referred to as model 1 or model 2, depending on whether the planning horizon is N or n. The aforementioned problem is model 1. For model 2 as well, we can find the similar alternate correspondence {p[m],m≥ 0}. We treat both the no-information model and the full-information model and examine the limiting behaviors of their optimal rules and optimal values related to the alternate correspondences as n→∞. A generalization of the no-information model is given. It is worth mentioning that the alternate correspondences for model 1 and model 2 are respectively related to the urn sampling models without replacement and with replacement.

Voting in collective best-choice problem with complete information

This paper considers a noncooperative m-player best-choice game with complete information about the quality parameters of incoming candidates. Collective decision-making is based on voting. The optimal threshold strategies and payoffs of the players are found depending on the voting threshold. The results of numerical simulation are presented.

A game-theoretic model of TV show “The Voice”

This paper proposes a game-theoretic model of the two-player best-choice problem with incomplete information. The players (experts) choose between objects by observing their quality in the form of two components forming a sequence of random variables (xi, yi), i = 1,..., n. By assumption, the first quality component xi is known to the players and the second one yi is hidden. A player accepts or declines an object based on the first quality component only. A player with the maximal sum of the components becomes the winner in the game. The optimal strategies are derived in the cases of independent and correlated quality components.

On the optimal stopping problems with monotone thresholds

As a class of optimal stopping problems with monotone thresholds, we define the candidate-choice problem (CCP) and derive two formulae for calculating its expected payoff. We apply the first formula to a particular CCP, i.e. the best-choice duration problem treated by Ferguson et al. (1992). The recall case is also examined as a comparison. We also derive the distribution of the stopping time of the CCP and find, as a by-product, that the best-choice problem has a remarkable feature in that the optimal probability of choosing the best is just the expected value of the (proportional) stopping time. The similarity between the best-choice duration problem and the best-choice problem with uniform freeze studied by Samuel-Cahn (1996) is recognized.

On the optimal stopping problems with monotone thresholds

As a class of optimal stopping problems with monotone thresholds, we define the candidate-choice problem (CCP) and derive two formulae for calculating its expected payoff. We apply the first formula to a particular CCP, i.e. the best-choice duration problem treated by Ferguson et al. (1992). The recall case is also examined as a comparison. We also derive the distribution of the stopping time of the CCP and find, as a by-product, that the best-choice problem has a remarkable feature in that the optimal probability of choosing the best is just the expected value of the (proportional) stopping time. The similarity between the best-choice duration problem and the best-choice problem with uniform freeze studied by Samuel-Cahn (1996) is recognized.

The full-information best-choice problem with uniform or gamma horizons

A decision-maker has to choose one from among a Poisson stream of i.i.d. bids, with no recall. The stream stops at a random time with a uniform (in the first case) or Erlang (in the second case) distribution. We solve the problem explicitly for maximal expected gain for bids that may take on any finite number of values. A fast procedure to solve the problem for fixed horizon is presented as well.

Bayes' Model of the Best-Choice Problem with Disorder

We consider the best-choice problem with disorder and imperfect observation. The decision-maker observes sequentially a known number of i.i.d random variables from a known distribution with the object of choosing the largest. At the random time the distribution law of observations is changed. The random variables cannot be perfectly observed. Each time a random variable is sampled the decision-maker is informed only whether it is greater than or less than some level specified by him. The decision-maker can choose at most one of the observation. The optimal rule is derived in the class of Bayes' strategies.

Optimal Choice of the Best Available Applicant in Full-Information Models

The problem we consider here is a full-information best-choice problem in which n applicants appear sequentially, but each applicant refuses an offer independently of other applicants with known fixed probability 0≤q<1. The objective is to maximize the probability of choosing the best available applicant. Two models are distinguished according to when the availability can be ascertained; the availability is ascertained just after the arrival of the applicant (Model 1), whereas the availability can be ascertained only when an offer is made (Model 2). For Model 1, we can obtain the explicit expressions for the optimal stopping rule and the optimal probability for a given n. A remarkable feature of this model is that, asymptotically (i.e. n→∞), the optimal probability becomes insensitive to q and approaches 0.580 164. The planar Poisson process (PPP) model provides more insight into this phenomenon. For Model 2, the optimal stopping rule depends on the past history in a complicated way and seems to be intractable. We have not solved this model for a finite n but derive, via the PPP approach, a lower bound on the asymptotically optimal probability.

Optimal Choice of the Best Available Applicant in Full-Information Models

The problem we consider here is a full-information best-choice problem in which n applicants appear sequentially, but each applicant refuses an offer independently of other applicants with known fixed probability 0≤q<1. The objective is to maximize the probability of choosing the best available applicant. Two models are distinguished according to when the availability can be ascertained; the availability is ascertained just after the arrival of the applicant (Model 1), whereas the availability can be ascertained only when an offer is made (Model 2). For Model 1, we can obtain the explicit expressions for the optimal stopping rule and the optimal probability for a given n. A remarkable feature of this model is that, asymptotically (i.e. n→∞), the optimal probability becomes insensitive to q and approaches 0.580 164. The planar Poisson process (PPP) model provides more insight into this phenomenon. For Model 2, the optimal stopping rule depends on the past history in a complicated way and seems to be intractable. We have not solved this model for a finite n but derive, via the PPP approach, a lower bound on the asymptotically optimal probability.

Optimal Stopping with Rank-Dependent Loss

For τ, a stopping rule adapted to a sequence ofnindependent and identically distributed observations, we define the loss to be E[q(Rτ)], whereRjis the rank of thejth observation andqis a nondecreasing function of the rank. This setting covers both the best-choice problem, withq(r) =1(r> 1), and Robbins' problem, withq(r) =r. Asntends to ∞, the stopping problem acquires a limiting form which is associated with the planar Poisson process. Inspecting the limit we establish bounds on the stopping value and reveal qualitative features of the optimal rule. In particular, we show that the complete history dependence persists in the limit; thus answering a question asked by Bruss (2005) in the context of Robbins' problem.

Optimal Stopping with Rank-Dependent Loss

For τ, a stopping rule adapted to a sequence of n independent and identically distributed observations, we define the loss to be E[q(Rτ)], where Rj is the rank of the jth observation and q is a nondecreasing function of the rank. This setting covers both the best-choice problem, with q(r) = 1(r > 1), and Robbins' problem, with q(r) = r. As n tends to ∞, the stopping problem acquires a limiting form which is associated with the planar Poisson process. Inspecting the limit we establish bounds on the stopping value and reveal qualitative features of the optimal rule. In particular, we show that the complete history dependence persists in the limit; thus answering a question asked by Bruss (2005) in the context of Robbins' problem.

Optimal Shopping When the Sales Are on—a Markovian Full-Information Best-Choice Problem

We study a full-information best-choice problem viewed in a shopping context. A certain commodity can be found at certain random times with stochastically fluctuating prices. While the prices may have a tendency to decrease, the instants at which items are offered become less frequent and it is possible that the item currently found will be the last one. The prospective customer's objective is to buy at the right time so as to minimize the expected price of the acquired item. We propose a two-dimensional Markov chain model with a rather general continuous-time point process structure and dependence of the random prices on the availability times of the items. The value function v of the associated optimal stopping problem is characterized as the smallest solution of a two-dimensional integral equation; this allows us to find the optimal policy under certain conditions. In particular, we consider a nonhomogeneous Poisson model for which more specific results can be obtained. We derive a differential equation of which v is the uniformly smallest nonnegative solution. This way v is determined up to a boundary condition at infinity. We provide criteria for identifying a solution as the value function and also for the natural stopping rule to be optimal. Several examples are given.

Winning Rate in the Full-Information Best-Choice Problem

Following a long-standing suggestion by Gilbert and Mosteller, we derive an explicit formula for the asymptotic winning rate in the full-information best-choice problem.