---
title: Ex Ante Free-Order Prophet Inequalities
url: https://www.emergentmind.com/topics/ex-ante-free-order-prophet-inequalities
type: topic
---

# Ex Ante Free-Order Prophet Inequalities

Searching arXiv for the cited and related prophet-inequality papers to ground the article and retrieve arXiv identifiers.
Ex ante free-order prophet inequalities occupy two closely related positions in the prophet-inequality literature. In the classical single-choice setting, “free order” or “order selection” means that the gambler may choose the arrival order of independent distributions before observing any realizations; the objective is a competitive ratio against the prophet’s expected offline maximum [2304.04024]. In the matroid setting, an “ex-ante prophet inequality” compares an online policy not to the offline optimum directly, but to an ex-ante relaxation defined over the matroid polytope; this viewpoint yields optimal online contention resolution schemes (OCRSs) and random-order contention resolution schemes (RCRSs) for matroids [1806.09251]. The two usages share the same ex-ante informational structure—distributions are known before values are realized—but differ in benchmark, feasibility constraints, and algorithmic consequences.

## 1. Conceptual scope and model distinctions

The classical prophet-inequality problem starts from independent nonnegative random variables \(V_1,\ldots,V_n\) drawn from known distributions \(D_1,\ldots,D_n\). A prophet observes all realizations and achieves
\[
E[\max_i V_i].
\]
An online gambler observes the values sequentially in some order \(\pi\), accepts one value irrevocably or continues, and aims to maximize
\[
E[V_{\pi_\tau}],
\]
where \(\tau\) is the stopping time. The central objective is the largest constant \(c\) such that
\[
E[V_{\pi_\tau}] / E[\max_i V_i] \ge c
\]
for every instance; \(c\) is the competitive ratio [2304.04024].

Within this framework, three order models are distinguished. In adversarial order, the order is fixed by an adversary, and the optimal ratio is \(1/2\). In random order, the permutation is drawn uniformly from \(S_n\). In order-selection (OS), also described as “ex-ante free-order,” the gambler chooses \(\pi \in S_n\) after seeing all distributions but before observing realizations; the data states that clearly OS \(\ge\) random-order \(\ge\) adversarial [2304.04024].

A separate but connected line of work concerns matroid prophets. There, the feasible set is an independent set of a matroid \(M=(N,\mathcal{I})\), and the online algorithm may accept multiple elements subject to independence. The distinguishing feature is the benchmark: instead of competing directly with the offline optimum, the algorithm is analyzed against an ex-ante relaxation over the matroid polytope \(P_M\) [1806.09251]. This distinction is essential. A common misconception is to identify ex-ante prophet inequalities with free-order prophet inequalities in the OS sense; the literature represented here uses “ex ante” in both senses, but the associated benchmarks and algorithmic structures are different.

| Setting | Order/benchmark structure | Best guarantee in the data |
|---|---|---|
| Classical adversarial-order | Offline maximum benchmark | \(1/2\) |
| Classical random-order | Offline maximum benchmark | hardness \(0.7235\); best algorithms \(\simeq 0.6697\) |
| Classical order-selection | Offline maximum benchmark, order chosen ex ante | \(0.7258\) |
| Matroid adversarial ex-ante | Ex-ante LP benchmark \(OPT_{\mathrm{ex}}\) | \(1/2\) |
| Matroid random-order ex-ante | Ex-ante LP benchmark \(OPT_{\mathrm{ex}}\) | \(1-1/e\) |

## 2. Ex-ante relaxation for matroid prophet inequalities

For matroid prophets, let \(M=(N,\mathcal{I})\) be a matroid with rank function \(r(\cdot)\), and let \(D_i\) be the known distribution of \(v_i \ge 0\) for each \(i \in N\). The matroid polytope is
\[
P_M = \{x \in [0,1]^N : \forall S \subseteq N,\ \sum_{i\in S} x_i \le r(S)\}.
\]
If \(F_i^{-1}(\cdot)\) denotes the inverse CDF of \(D_i\), define
\[
y_i := E[v_i \mid v_i \text{ is in its top } x_i\text{–quantile}],
\]
so that
\[
\int_0^{x_i} F_i^{-1}(1-t)\,dt = x_i \cdot y_i.
\]
The ex-ante relaxation is the convex program
\[
\max_{x\in[0,1]^N} \sum_{i=1}^n x_i y_i
\quad\text{subject to}\quad
\forall S\subseteq N:\ \sum_{i\in S} x_i \le r(S),
\]
or equivalently \(x \in P_M\). Its optimum is denoted
\[
OPT_{\mathrm{ex}} := \max_{x\in P_M} \sum_i x_i y_i.
\]
This objective is an upper bound on the expected offline optimum [1806.09251].

This benchmark isolates the ex-ante structure of the instance before any realizations occur. The resulting optimization problem is deterministic and convex, and it compresses the stochastic instance into quantile masses \(x_i\) and conditional top-quantile values \(y_i\). A plausible implication is that this formulation is especially well suited to reductions to rounding problems, because the decision variables already live in the matroid polytope.

## 3. Adversarial-order ex-ante prophet inequality and the optimal \(1/2\)-OCRS

The adversarial-order theorem states: for any matroid \(M\) and independent distributions \((D_i)_{i\in N}\), there is a deterministic online algorithm in adversarial arrival order whose expected reward is at least
\[
\frac12 \cdot OPT_{\mathrm{ex}}.
\]
The proof first reduces to the case where each \(v_i\) is Bernoulli taking value \(y_i\) with probability \(x_i\). It then considers a correlated ex-ante value vector obtained by sampling an independent set \(I\) from an \(x\)-decomposition of \(x \in P_M\) and setting \(\hat v_i = y_i\) if \(i \in I\), and \(0\) otherwise [1806.09251].

The algorithm maintains the accepted set \(A \subseteq N\). If \(A_k\) denotes the set of the first \(k\) arrivals accepted, it defines the residual function
\[
r(k) := E[R(A_k,\hat v)],
\]
where \(R(A,\hat v)\) is the value of the maximum \(\hat v\)-weight independent set in the contraction \(M/A\). With the uniform threshold parameter \(\alpha = 1/2\), the algorithm accepts element \(i\) at state \(A\) if and only if \(A \cup \{i\} \in \mathcal{I}\) and
\[
\hat v_i \ge \tau_i := \frac12\, b_i(A),
\]
where
\[
b_i(A) := E[R(A,\hat v)-R(A\cup\{i\},\hat v)]
\]
is the base-price of \(i\) at state \(A\) [1806.09251].

The analysis decomposes the reward into
\[
\mathrm{Revenue} = \sum_{i\in A} \frac12\, b_i(A_{i-1}),
\qquad
\mathrm{Utility} = \sum_{i\in A} \bigl(\hat v_i - \tfrac12 b_i(A_{i-1})\bigr)^+.
\]
Two telescoping arguments show
\[
E[\mathrm{Revenue}] = \frac12\,(r(0)-r(n)),
\qquad
E[\mathrm{Utility}] \ge \frac12\, r(n),
\]
and hence \(E[\mathrm{Total}] \ge \frac12\, r(0)=\frac12\, OPT_{\mathrm{ex}}\) [1806.09251].

This ex-ante prophet inequality is then converted into an OCRS by a standard dual-fitting argument. For online rounding of \(x \in P_M\), the preprocessing computes an optimal \(x\) and the corresponding \(y_i\), and then uses the same residual base-prices \(b_i(A)\). Online, element \(i\) becomes active with probability \(x_i\); if active and \(A \cup \{i\} \in \mathcal{I}\) and \(y_i \ge \frac12 b_i(A)\), it is accepted, otherwise rejected. The resulting randomized contention resolution scheme selects each \(i\) with probability at least \(\frac12 x_i\) while always maintaining \(A \in \mathcal{I}\), hence it is a \(1/2\)-OCRS [1806.09251].

## 4. Random-order ex-ante prophet inequalities and optimal \((1-1/e)\)-RCRSs

In the random-order prophet model for matroids, values are drawn independently at time \(0\), the adversary may see \(x\), but the permutation of arrivals is uniformly random. On observing \(v_i\) and its time \(t_i \in [0,1]\), the algorithm must decide irrevocably. Under the same ex-ante benchmark \(OPT_{\mathrm{ex}}\), there is an online policy in random arrival order with expected reward at least
\[
(1-1/e)\cdot OPT_{\mathrm{ex}}.
\]
This is the random-order ex-ante prophet inequality for matroids [1806.09251].

As in the adversarial case, the proof reduces to Bernoulli \(\hat v\)-values. Let \(A_t\) denote the set of accepted elements arriving before time \(t\), and define
\[
r(t) := E[R(A_t,\hat v)].
\]
The algorithm uses the time-varying threshold
\[
\alpha(t) := 1-\exp(t-1), \qquad t\in[0,1].
\]
When element \(i\) arrives at time \(t\) with \(\hat v_i \in \{0,y_i\}\), it is accepted if \(A_t \cup \{i\} \in \mathcal{I}\) and
\[
\hat v_i \ge \alpha(t)\, b_i(A_t).
\]
The reward is again split into revenue and utility, with
\[
E[\mathrm{Revenue}] = -\int_0^1 \alpha(t)\, r'(t)\, dt,
\qquad
E[\mathrm{Utility}] \ge \int_0^1 (1-\alpha(t))\, r(t)\, dt.
\]
Because \(\alpha\) satisfies \(1-\alpha(t)+\alpha'(t)=0\), the total reward obeys
\[
E[\mathrm{Total}] \ge [-r(t)\alpha(t)]_0^1 = (1-1/e)\, r(0),
\]
which equals \((1-1/e)\, OPT_{\mathrm{ex}}\) [1806.09251].

The induced rounding scheme is an RCRS, and the guarantee is tight. The data states that tightness for rank \(1\) shows that no \(c\)-OCRS in adversarial order can do better than \(c=1/2\), and no \(c\)-RCRS in random order can do better than \(c=1-1/e\). The exchange lemma,
\[
\sum_{i\in S} b_i(A) \le R(A),
\]
holds because the \(b_i(\cdot)\) are the critical values of the usual greedy algorithm in the contracted matroid, and this lemma underpins both the adversarial and random-order analyses [1806.09251].

## 5. Classical free-order prophet inequalities: order selection

In the classical single-choice model, order-selection is the setting in which the gambler chooses the arrival order using the distributions alone, then observes the realized values in that order. The data explicitly identifies this as the “Order-Selection (OS, also ‘ex-ante free-order’)” model, and notes that one only needs to consider nonrandomized stopping times \(\tau\) by standard arguments [2304.04024].

Historically, Hill (1983) showed that by choosing the order one can break the \(1/2\) barrier of adversarial order. Chawla–Hartline et al. (2010) gave a simple posted-price style single-threshold algorithm attaining \(0.6321\). Later refinements increased the ratio to \(0.6346\), \(0.6541\), and \(0.6697\), with the latter associated in the data with blind strategies of Correa–Saona–Ziliotto (2021). Peng–Tang (FOCS 2022) introduced a continuous-time arrival-design with a one-parameter threshold function and proved a \(0.7251\)-competitive ratio for OS. Bubna–Chiplunkar (EC 2023) refined the threshold-optimization analysis to obtain \(0.7258\) [2304.04024].

The Peng–Tang scheme embeds the \(n\) variables on \([0,1]\) by selecting arrival times \(t_1 < \cdots < t_n\), one per distribution, so that the conditional distribution of the maximum is spread out in time; equivalently, one chooses an absolutely continuous order-measure on \(S_n\). A threshold function \(g(t)>0\) is then associated to each time \(t\), and the gambler accepts the first arriving value \(V_i\) with \(V_i \ge g(t)\). The function \(g\) is chosen by solving an integral equation or differential equation arising from continuous-time backward induction [2304.04024].

The refinement by Bubna–Chiplunkar splits \([0,1]\) into more segments and solves a small convex program for the worst-case instance. The data states that the improvement uses analytic lower bounds on the worst-case ratio and explicitly bounds the integrals arising in the backward-induction analysis, thereby avoiding brute-force simulation. The outcome is an explicit family of thresholds, given in closed form up to a few parameters, guaranteeing ratio at least \(0.7258\) on every instance [2304.04024].

## 6. Separation from random order, methodological themes, and open questions

A central recent development is the rigorous separation between order selection and random order. For random order, the data reports a new hardness result of \(0.7235\): no algorithm can achieve a larger ratio in the random-order model. The proof uses an explicit “three-point-plus-constant” instance consisting of \(n\) i.i.d. random variables taking \(\{n\) with probability \(1/n^2\), \(b\) with \(p/n\), \(0\) else\}\) and one extra constant \(a\), with parameters \(a \approx 0.789\), \(b \approx 1.24\), and \(p \approx 0.421\). By asymptotically analyzing exact backward-induction thresholds as \(n \to \infty\), the gambler-to-prophet ratio is upper-bounded by \(0.7235 + o(1)\) [2304.04024].

This establishes that order selection strictly improves over simply taking a uniformly random order, because OS admits a \(0.7258\)-competitive algorithm while the random-order model has hardness at most \(0.7235\). The separation is significant precisely because random order and ex-ante order choice are sometimes conflated. The data makes the distinction explicit: carefully choosing the order, instead of simply taking it at random, benefits the gambler [2304.04024].

Across both the matroid and single-choice lines of work, a common methodological pattern is visible. In the matroid setting, ex-ante relaxations over \(P_M\), Bernoulli reductions, residual-value processes, and duality connect prophet inequalities to contention resolution [1806.09251]. In the single-choice free-order setting, continuous-time relaxations, backward induction, and threshold functions \(g(t)\) organize both algorithm design and lower-bound analysis [2304.04024]. This suggests a broader unifying principle: ex-ante structure is exploited either through relaxations over feasible polytopes or through ex-ante control of temporal placement.

The open questions stated in the data remain substantial. For classical OS, the optimal competitive ratio is unknown; the IID special case has tight ratio \(\simeq 0.7451\), leaving a gap of roughly \(0.02\). The data also asks whether full OS reaches \(0.7451\), whether there are extremal instances forcing OS ratio below \(0.7451\), and whether richer free-order choices can lower or characterize the optimum more sharply. For combinatorial prophet inequalities, the open directions include extending the \(1/2\)-OCRS guarantee to the fully online adversary of [FSZ16] and determining whether similar optimal ex-ante free-order prophet inequalities exist for richer constraints such as intersections of matroids [1806.09251]. Together, these questions place ex ante free-order prophet inequalities at the intersection of optimal stopping, stochastic combinatorial optimization, and online rounding.

Source: https://www.emergentmind.com/topics/ex-ante-free-order-prophet-inequalities