國立虎尾科技大學 |

Sequential Bilevel Linear Programming with Incomplete Information and Learning.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Sequential Bilevel Linear Programming with Incomplete Information and Learning./
Author:	Borrero, Juan Sebastian.
Description:	1 online resource (179 pages)
Notes:	Source: Dissertation Abstracts International, Volume: 79-04(E), Section: B.
Contained By:	Dissertation Abstracts International79-04B(E).
Subject:	Operations research. -
Online resource:	click for full text (PQDT)
ISBN:	9780355409642

Sequential Bilevel Linear Programming with Incomplete Information and Learning.
Borrero, Juan Sebastian.

Sequential Bilevel Linear Programming with Incomplete Information and Learning. - 1 online resource (179 pages)

Source: Dissertation Abstracts International, Volume: 79-04(E), Section: B.

Thesis (Ph.D.)

Includes bibliographical references

We present a framework for a class of sequential decision-making problems in the context of bilevel linear programming, where a leader and a follower repeatedly interact. At each period, the leader allocates resources that can modify the performance of the follower (e.g., as in interdiction or defender-attacker problems). The follower, in turn, optimizes some cost function over a set of activities that depends on the leader's decision. While the follower has complete knowledge of his problem, the leader, who decides as to optimize her objective function, has only partial information. As such, she needs to learn about the cost parameters, available resources, and the follower's activities from the feedback generated by the follower's actions. We measure the performance of any given leader's decision-making policy in terms of its time-stability, defined as the number of periods it takes the policy to match the actions of an oracle decision-maker with complete information of the bilevel problem.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9780355409642Subjects--Topical Terms:

573517
Operations research.
Index Terms--Genre/Form:

554714
Electronic books.

Sequential Bilevel Linear Programming with Incomplete Information and Learning.
LDR:03637ntm a2200361Ki 4500 001 910859
005 20180517112612.5
006 m o u
007 cr mn||||a|a||
008 190606s2017 xx obm 000 0 eng d
020 $a 9780355409642
035 $a (MiAaPQ)AAI10692491
035 $a AAI10692491
040 $a MiAaPQ $b eng $c MiAaPQ
099 $a TUL $f hyy $c available through World Wide Web
100 1 $a Borrero, Juan Sebastian. $3 1182343
245 1 0 $a Sequential Bilevel Linear Programming with Incomplete Information and Learning.
264 0 $c 2017
300 $a 1 online resource (179 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 79-04(E), Section: B.
500 $a Adviser: Oleg A. Prokopyev.
502 $a Thesis (Ph.D.) $c University of Pittsburgh $d 2017.
504 $a Includes bibliographical references
520 $a We present a framework for a class of sequential decision-making problems in the context of bilevel linear programming, where a leader and a follower repeatedly interact. At each period, the leader allocates resources that can modify the performance of the follower (e.g., as in interdiction or defender-attacker problems). The follower, in turn, optimizes some cost function over a set of activities that depends on the leader's decision. While the follower has complete knowledge of his problem, the leader, who decides as to optimize her objective function, has only partial information. As such, she needs to learn about the cost parameters, available resources, and the follower's activities from the feedback generated by the follower's actions. We measure the performance of any given leader's decision-making policy in terms of its time-stability, defined as the number of periods it takes the policy to match the actions of an oracle decision-maker with complete information of the bilevel problem.
520 $a Three types of bilevel models are considered: Shortest path interdiction, max-min bilevel linear problems, and asymmetric bilevel linear problems. For shortest path interdiction we discuss greedy and pessimistic policies, and show that their time stability is upper-bounded by the number of arcs in the network; moreover, these policies are not dominated by any nongreedy or non-pessimistic policy. We refine these ideas into the more general max-min bilevel problems. Here we show that there is a class of greedy and robust policies that have the best possible worst-case performance, eventually match the oracle's actions, provide a real-time optimality certificate, and can be computed using mixed-integer linear programming. These policies, however, do not retain their features for asymmetric bilevel problems. For this setting we study the performance of greedy and best-case policies and show that they keep many of the attractive properties that the greedy and robust policies have for the max-min case.
520 $a By performing computational experiments under different configurations, we show that the proposed policies compare favorably against different benchmark policies. Moreover, they perform reasonably close to the semi-oracle, that is a novel decision-maker we introduce that provides a lower bound on the time-stability of any policy.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Operations research. $3 573517
650 4 $a Industrial engineering. $3 679492
655 7 $a Electronic books. $2 local $3 554714
690 $a 0796
690 $a 0546
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of Pittsburgh. $3 1178873
773 0 $t Dissertation Abstracts International $g 79-04B(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10692491 $z click for full text (PQDT)