an online allocation problem: dark pools
TRANSCRIPT
![Page 1: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/1.jpg)
An Online Allocation Problem: Dark Pools
Peter BartlettStatistics and EECS
UC Berkeley
Joint work with Alekh Agarwal and Max Dama.
slides at http://www.stat.berkeley.edu/∼bartlett
![Page 2: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/2.jpg)
Prediction in Probabilistic Settings
I i.i.d. (X ,Y ), (X1,Y1), . . . , (Xn,Yn) from X × Y(e.g., Y is preference score vector).
I Use data (X1,Y1), . . . , (Xn,Yn) to choose fn : X → A withsmall risk,
R(fn) = E`(Y , fn(X )).
![Page 3: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/3.jpg)
Online Learning
I Repeated game:
Player chooses at
Adversary reveals `t
I Example: `t (at ) = loss(yt ,at (xt )).
I Aim: minimize∑
t
`t (at ), compared to the best
(in retrospect) from some class:
regret =∑
t
`t (at )−mina∈A
∑t
`t (a).
I Data can be adversarially chosen.
![Page 4: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/4.jpg)
Online Learning: Motivations
1. Adversarial model is appropriate forI Computer security.I Computational finance.
2. Understanding statistical prediction methods.3. Online algorithms are also effective in probabilistic settings.
![Page 5: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/5.jpg)
The Dark Pools Problem
I Computational finance: adversarial setting is appropriate.I Online algorithm improves on best known algorithm for
probabilistic setting.
![Page 6: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/6.jpg)
Dark Pools
Instinet,Chi-X,Knight Match, ...
International Securities Exchange,Investment Technology Group(POSIT),
I Crossing networks.I Alternative to open exchanges.I Avoid market impact by hiding transaction size and traders’
identities.
![Page 7: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/7.jpg)
Dark Pools
![Page 8: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/8.jpg)
Dark Pools
![Page 9: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/9.jpg)
Dark Pools
![Page 10: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/10.jpg)
Dark Pools
![Page 11: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/11.jpg)
Allocations for Dark Pools
The problem: Allocate orders to several dark poolsso as to maximize the volume of transactions.
I Volume V t must be allocated across K venues: v t1, . . . , v
tK ,
such that∑K
k=1 v tk = V t .
I Venue k can accommodate up to stk , transacts
r tk = min(v t
k , stk ).
I The aim is to maximizeT∑
t=1
K∑k=1
r tk .
![Page 12: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/12.jpg)
Allocations for Dark Pools
I Allocation v t1, . . . , v
tK ranks the K venues.
I Loss is not discrete: it is summed across venues, anddepends on the allocations in a piecewise-linear, convex,monotone way.
![Page 13: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/13.jpg)
Allocations for Dark Pools: Probabilistic Assumptions
Previous work: (Ganchev, Kearns, Nevmyvaka and Wortman, 2008)
I Assume venue volumes are i.i.d.:st
k , k = 1, . . . ,K , t = 1, . . . ,T.I In deciding how to allocate the first unit,
choose the venue k where Pr(stk > 0) is largest.
I Allocate the second and subsequent units in decreasingorder of venue tail probabilities.
I Algorithm: estimate the tail probabilities (Kaplan-Meierestimator—data is censored), and allocate as if theestimates are correct.
![Page 14: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/14.jpg)
Allocations for Dark Pools: Adversarial Assumptions
I.i.d. is questionable:I one party’s gain is another’s lossI volume available now affects volume remaining in futureI volume available at one venue affects volume available at
othersIn the adversarial setting, we allow an arbitrary sequence ofvenue capacities (st
k ), and of total volume to be allocated (V t ).The aim is to compete with any fixed allocation order.
![Page 15: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/15.jpg)
Continuous Allocations
We wish to maximize a sum of (unknown) concave functions ofthe allocations:
J(v) =T∑
t=1
K∑k=1
min(v tk , s
tk ),
subject to the constraint∑K
k=1 v tk ≤ V t .
The allocations are parameterized as distributions over the Kvenues:
x1t , x
2t , . . . ∈ ∆K−1 = (K − 1)-simplex.
Here, x1t determines how the first unit is allocated, x2
t thesecond, ...
The algorithm allocates to the k th venue: v tk =
V t∑v=1
xvt ,k .
![Page 16: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/16.jpg)
Continuous Allocations
We wish to maximize a sum of (unknown) concave functions ofthe distributions:
J =T∑
t=1
K∑k=1
min(v tk (xv
t ,k ), stk ).
Want small regret with respect to an arbitrary distribution xv ,and hence w.r.t. an arbitrary allocation.
regret =T∑
t=1
K∑k=1
min(v tk (xv
k ), stk )− J.
![Page 17: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/17.jpg)
Continuous Allocations
We use an exponentiated gradient algorithm:
Initialize xv1,i = 1
K for v = 1, . . . ,V.for t = 1, . . . ,T do
Set v tk =
∑V T
v=1 xvt ,k .
Receive r tk = minv t
k , stk.
Set gvt ,k = ∇xv
t,kJ.
Update xvt+1,k ∝ xv
t ,k exp(ηgvt ,k ).
end for
![Page 18: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/18.jpg)
Continuous Allocations
Theorem: For all choices of V t ≤ V and of stk , ExpGrad has
regret no more than 3V√
T ln K .
![Page 19: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/19.jpg)
Continuous Allocations
Theorem: For all choices of V t ≤ V and of stk , ExpGrad has
regret no more than 3V√
T ln K .
Theorem: For every algorithm, there are sequences V t and stk
such that regret is at least V√
T ln K/16.
![Page 20: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/20.jpg)
Experimental results
0 200 400 600 800 1000 1200 1400 1600 1800 20000
0.5
1
1.5
2
2.5
3
3.5
4x 106
Round
Cum
ulat
ive
Rew
ard
Cumulative Reward at Each Round
Exp3ExpGradOptKMParML
![Page 21: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/21.jpg)
Continuous Allocations: i.i.d. data
I Simple online-to-batch conversions show ExpGrad obtainsper-trial utility within O(T−1/2) of optimal.
I Ganchev et al bounds:per-trial utility within O(T−1/4) of optimal.
![Page 22: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/22.jpg)
Discrete allocations
I Trades occur in quantized parcels.I Hence, we cannot allocate arbitrary values.I This is analogous to a multi-arm bandit problem:
I We cannot directly obtain the gradient at the current x .I But, we can estimate it using importance sampling ideas.
Theorem: There is an algorithm for discrete allocation with ex-pected regret O((VTK )2/3).Any algorithm has regret Ω((VTK )1/2).
![Page 23: An Online Allocation Problem: Dark Pools](https://reader034.vdocuments.us/reader034/viewer/2022052601/628ca9e59768b70c65120a12/html5/thumbnails/23.jpg)
Dark Pools
I Allow adversarial choice of volumes and transactions.I Per trial regret rate superior to previous best known
bounds for probabilistic setting.I In simulations, performance comparable to (correct)
parametric model’s, and superior to nonparametricestimate.