This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ting the value for all possible bundles
Par99 . In comparison, an OR bidding language S1 ; p1 or S2 ; p2 , which states that the
agent wants S1 or S2 or both, has a linearspace representation of this valuation function.
Nisan observes that other combinations, such as XORofOR languages and ORofXOR languages, allow compact representations of certain preference structures and make
tradeo s across expressiveness and compactness. Nisan proposes an OR* bidding language,
which is expressive enough to be able to represent arbitrary preferences over discrete items,
and as compact a representation as both ORofXOR and XORofOR representations.
However, Nisan provides an example with no polynomialsize representation even with the
OR* language.
The expressiveness of a bidding language, or the compactness of representations that it
permits, becomes even more important when one considers the agent's underlying valuation
problem.
Suppose that an agent must solve an NPhard constrained optimization problem P
to compute its value for a set of items, with objective function g and constraints C . In the
74 XOR representation the agent must solve this problem P once for every possible input
S G , i.e. requiring an exponential number of solutions to an NPhard problem. Now
consider an alternative bidding language, that allows the agent to send the speci cation of
its optimization problem, i.e. P = g; C directly to the auctioneer. Strategyproofness
is not a ected assuming the agent can trust the mechanism to interpret this bidding
language faithfully, but the agent saves a lot of value computation.
In general, we might consider a language in which the agent can send a bidding
program" to the auctioneer, that the auctioneer will then execute as necessary to compute
an agent's value for di erent subsets of items Nis00 . This is really just the extreme
limit of the revelation principle: rather than requiring an agent to solve its local problem
and compute its value for all possible outcomes, simply allow the agent to send the local
problem speci cation directly to the auctioneer.
From the perspective of the bidding agent this approach simpli es its valuation problem
whenever the speci cation of its local problem is simpler than actually computing its value
for all possible outcomes. A bidding program allows an agent to feed that speci cation
directly to the auctioneer.
From the perspective of the auctioneer, this is an even more centralized solution than
providing a complete valuation function, and has worsestill privacy implications. The
bidding program approach shifts the valuation computational burden from agents to the
auctioneer. Notice for example that if the bidding program provides only black box"
functionality, e.g. b : 2G ! R, the mechanism must compute bS for all S G
unless other consistency rules such as free disposal apply to an agents' values to compute
the e cient solution. However, if the bidding program, or language, provides a richer
functionality for example allowing e cient pruning the value bS 0 on all bundles S 0 S
is less than bS "; or computing approximate...
View
Full
Document
This document was uploaded on 03/19/2014 for the course COMP CS286r at Harvard.
 Fall '13
 DavidParkes

Click to edit the document details