Expense constrained bidder optimization in repeated auctions.
the nature of what this budget limit means for the bidders themselves is somewhat of a mystery. There seems to be some risk control element to it, some purely administrative element to it, some bounded-rationality element to it, and more...
A two parameter model for expense constraints in online budgeting problems.
Optimal bid can be mapped to static auction with a shaded virtual valuation.
Paper has more contents: MFE analysis and a finite horizon model.
6. Modeling Expense Constraints
“…the nature of what this budget limit means for the
bidders themselves is somewhat of a mystery. There
seems to be some risk control element to it, some
purely administrative element to it, some bounded-
rationality element to it, and more…”
-- “Theory research at google”, SIGACT News, 2008.
8. Responsibility for expense constraints
Auctioneer Bidder
Bids fixed -- Auction entry
throttled.
Bids adjusted dynamically.
Online bipartite matching
between queries and bidders.
Online knapsack type problems.
Expense constraints
= fixed budget.
Possible to model more general
expense constraints.
10. Preview
Sequential X-auction with true value v
≡
Static X-auction with virtual value: shade* v
X can be SP, GSP, FP, etc. (any quasi linear utility)
Shade(remaining balance B) =
𝑉(𝐵) will be characterized explicitly.
1
1 '( )V B
12. Overview
• Introduction
• Budgeted Second Price auctions
• A General Online Budgeting Framework
• Optimal Bids for Micro-Value Auctions
• Conclusion
13. Model: Budgeted Second Price
• Discrete time, indexed 𝒕 = 𝟎, 𝟏, 𝟐 …
• Balance: 𝒃(𝒕)
• Constant income per time slot - 𝒂 ≥ 𝟎
• I.I.D. environment sampled from
– Private valuation ~ 𝒗 (observable)
– Competing bid ~ 𝒑 (not observable)
• Decision variable is bid at time 𝑖: 𝒖 𝒕
– Can depend on 𝒗 𝒕 and 𝒃(𝒕), but not 𝒑(𝒕)
15. The Value Function
𝒗 𝜷 (𝒃) = 𝒔𝒖𝒑
𝓤 𝒕=𝟎
∞
𝒆−𝜷𝒕
𝔼[𝒈 𝒕 ] | 𝒃 𝟎 = 𝒃
• 𝑣 𝛽 (𝑏) : max utility starting with balance 𝑏
• Can use dynamic programming (“one step look
ahead”) to write out a functional fixed point
relation.
16. The Value Function
1 2
,
( ) max E 1{ } 1{ }
v pu b
v b u p T u p T
( )v p e v b a p
( )e v b a
But boundary conditions can not be inferred from the
DP argument.
Current
auction
1T
2T
17. Future opportunity cost
Characterization of value function
“Effective price” for nominal 𝒑 at balance 𝒃:
Theorem: Optimal bid is 𝒖*:
i.e: Buy all auctions with “effective price” ≤ 𝑣
𝑣 𝛽(𝑏) is a functional fixed point to:
( , )*u b v
,
( ) ( ) ( , )
v p
v b e v b a v p b
( , ) ( ( ) ( ))p b p e v b a v b a p
18. 1
,
( ) ( ) ( , )i i i
v p
v b e v b a v p b
Value Iteration:
𝛽 = 0.1
Each auction has miniscule utility
compared to overall utility: 𝛽 ≈ 0
19. Value Iteration:
𝛽 = 0.01
1
,
( ) ( ) ( , )i i i
v p
v b e v b a v p b
20. Numerical estimation when 𝛽 is small:
• State space quantization errors propagate due
to lack of boundary value.
• Need longer iterations over larger state space.
𝛽 ⟶ 0 will be studied under scaling:
( ) ( ) ( / )V B v b v B
Limiting case: micro-value auctions
21. Overview
• Introduction
• Budgeted Second Price Auctions
• A General Online Budgeting Framework
• Optimal Bids for Micro-Value Auctions
• Conclusion
24. Ex2: GSP Auction
Random environment:
𝜉 = 𝑣, 𝑝1 , … , 𝑝L , 𝛾1, … , 𝛾 𝐿
Observable part: ℱ0 = 𝜎 𝑣
Action: 𝑢 is the bid
Utility function:
Payment function:
1
1
( , ) 1{ } ( )
L
l l l l
l
g u p u p v p
1
1
( , ) 1{ }
L
l l l l
l
c u p u p p
Click events for L slots
25. Overview
• Introduction
• Budgeted Second Price Auctions
• A General Online Budgeting Framework
• Optimal Bids for Micro-Value Auctions
• Conclusion
26. Limiting Regime: 𝛽 ⟶ 0
( ) ( ) ( / )V B v b v B
Notation:
(( ]) [ , )E g ug u
(( ]) [ , )E c uc u
27. 𝑓(. ) is an inverse
and
is the minimum of:
Theorem
*( ), (0) ,
dV
f V V
dB
( )x
𝑉 𝐵 = lim
𝛽→0
𝑉𝛽 𝐵 is the solution to:
*
28. *( ') , (0) ,V V V
0
( ) sup( ( ) ( ) )
u
x ax g u c u x
F
* 0min ( )x x
𝑥
𝑉
𝑉’ 𝐵
𝑉(𝐵)
Theorem
( )x
*
30. Second Price Auction Example
Opponents bid p
Private Valuation 𝑣
𝜙(𝑥)
Value functions
31. Optimal bid
0 ( )
( )
sup( ( ) ( ) '( )) sup 1{ }( (1 '( ))
sup 1{ }
1 '( )
u u v
u v
g u c u V B uE p v p V B
v
u p p
V
E
B
F
i.e., Static SP with shaded valuation:
1 '( )
v
V B
𝒖* at balance B solves:
33. Optimal Bid: GSP
0
1
( ) 1
sup( ( ) ( ) '( ))
sup 1{ }
1 '( )
u
L
l l l l
u v l
g u c u V B
v
p u p p
V
E
B
F
Static GSP with “virtual valuation”:
1 '( )
v
V B
34. Proof Overview
• Variant: Retire with payoff 𝜂 when 𝑏 𝑡 = 0.
• Value function of variant converges to ODE
with initial value 𝜂.
• But what is the right boundary condition 𝜂?
To prove: lim sup 𝑉𝛽 0 ≤ 𝜂∗
≤ lim inf 𝑉𝛽 (0)
Because exit payoff ≈ optional Next 2 slides
35. Goal: Exhibit a sequence of policies parametrized by 𝛽 which can
achieve a scaled payoff 𝜂∗
as 𝛽 ⟶ 0
Lemma: For any ε > 0, there is a policy 𝑢* such that
𝑔 𝑢∗ > 𝜂∗
− ε AND 𝑐(𝑢∗) ≤ 𝑎
If 𝑢∗ could be played continuously, we can get arbitrarily
close to 𝜂∗
!
But every now and then balance is exhausted, so we need
a variant of u* that still manages to achieve nearly as
much payoff
𝜂∗ ≤ lim inf 𝑉𝛽 (0)
36. time
B(t)
B
Play U*
𝜂∗ ≤ lim inf 𝑉𝛽 (0)
Show that fraction of time spent in green phase by the random
walk gets arbitrarily close to 1 as 𝛽->0
37. Overview
• Introduction
• MDP for budgeted SP auctions
• A General Online Budgeting Framework
• Optimal Bids for Micro-Value Auctions
• Conclusion
38. Conclusion
• A two parameter model for expense
constraints in online budgeting problems.
• Optimal bid can be mapped to static auction
with a shaded virtual valuation.
• Paper has more contents: MFE analysis and a
finite horizon model.
Editor's Notes
Fundamental qn: why bother with additional constraints when utility function we defined is already taking into account the expense.
So maybe they exist, but is the effect significant?
Fundamental qn: why bother with additional constraints when utility function we defined is already taking into account the expense.
So maybe they exist, but is the effect significant?
Mention the two types of expense constraints
Discount factor interpretation – life time of bidder in the system.
Mention parametrization in terms of beta.
Well defined quantity, but to solve for it we need to optimize over a huge class of policies.
In a sequential setting , when an agent wins an auction, he pays a price equal to b’. But this is actually costlier than b’ in effect. How much costlier is it?
So lambda is a function that magnifies the opponent’s bid as a function of the current balance.
Mention value iteration interpretation to compute the function numerically.
Beta to zero implies the value function is blowing up to infty. But we need to scale the units correctly in terms of beta to understand the right behavior.
Phi maps rate of change of value function to the value itself.