NOTE: This item is not available outside the Texas A&M University network. Texas A&M affiliated users who are off campus can access the item through NetID and password authentication or by using TAMU VPN. Non-affiliated individuals should request a copy through their local library's interlibrary loan service.
An experimental design approach to evaluating multi-option algorithms illustrated on a new integer programming procedure
dc.contributor.advisor | Sielken, Robert J., Jr. | |
dc.creator | Riley, William James | |
dc.date.accessioned | 2020-08-21T22:03:25Z | |
dc.date.available | 2020-08-21T22:03:25Z | |
dc.date.issued | 1981 | |
dc.identifier.uri | https://hdl.handle.net/1969.1/DISSERTATIONS-644781 | |
dc.description | Typescript (photocopy). | en |
dc.description.abstract | Frequently algorithm users can select their solution strategy by choosing from among various options for each of several algorithm factors. If the algorithm will always eventually find a solution, the important question is which combination of options is likely to be "best". A general statistical approach to answering the question is illustrated in the context of a new integer linear programming algorithm where "best" is quickest. The integer programming algorithm is a sophisticated implicit enumeration algorithm. The four factors where the user must select an option are (1)augmenting partial solutions, (2)backtracking, (3)fathoming on the basis of binary feasibility and optimality indicators, and (4)use of linear programming on the relaxed problem which includes penalties, cuts, surrogate constraints, and associated fathoming. There are several options per factor so that the algorithm can function in over 14,000 different modes. A significant evaluation of the average effects of each option on the algorithm's speed and the interactions between options is obtained using analysis of variance techniques. The design of the experiment, the linear model, and the analysis of the resulting data are discussed. The generality of this approach to analyzing algorithm components is emphasized. | en |
dc.format.extent | x, 105 leaves | en |
dc.format.medium | electronic | en |
dc.format.mimetype | application/pdf | |
dc.language.iso | eng | |
dc.rights | This thesis was part of a retrospective digitization project authorized by the Texas A&M University Libraries. Copyright remains vested with the author(s). It is the user's responsibility to secure permission from the copyright holder(s) for re-use of the work beyond the provision of Fair Use. | en |
dc.rights.uri | http://rightsstatements.org/vocab/InC/1.0/ | |
dc.subject | Statistics | en |
dc.subject.classification | 1981 Dissertation R573 | |
dc.subject.lcsh | Mathematical statistics | en |
dc.subject.lcsh | Algorithms | en |
dc.subject.lcsh | Programming (Mathematics) | en |
dc.subject.lcsh | Experimental design | en |
dc.title | An experimental design approach to evaluating multi-option algorithms illustrated on a new integer programming procedure | en |
dc.type | Thesis | en |
thesis.degree.discipline | Philosophy | en |
thesis.degree.grantor | Texas A&M University | en |
thesis.degree.name | Doctor of Philosophy | en |
thesis.degree.name | Ph. D. in Philosophy | en |
thesis.degree.level | Doctorial | en |
dc.type.genre | dissertations | en |
dc.type.material | text | en |
dc.format.digitalOrigin | reformatted digital | en |
dc.publisher.digital | Texas A&M University. Libraries | |
dc.identifier.oclc | 7929383 |
Files in this item
This item appears in the following Collection(s)
-
Digitized Theses and Dissertations (1922–2004)
Texas A&M University Theses and Dissertations (1922–2004)
Request Open Access
This item and its contents are restricted. If this is your thesis or dissertation, you can make it open-access. This will allow all visitors to view the contents of the thesis.