Type of Document	Master's Thesis
Author	Khan, Tanvir
Author's Email Address	tkhan1@lsu.edu
URN	etd-04022007-211811
Title	Optimization for Software Release and Crash
Degree	Master of Science in Engineering Science (M.S.E.S.)
Department	Engineering Science (Interdepartmental Program)
Advisory Committee	Advisor Name Title Xiaoyue Jiang Committee Chair Dan B. Rinks Committee Member Junfang Yu Committee Member
Keywords	de-eutrophication uniform parameter optimal crash policy crash optimal stopping Software Reliability Growth Model optimal release policy release time crash time stochastic comparison reliability cost structure monotonicity optimal policy release non-homogeneous Poisson process jumping process NHPP probability survival function control-limit policy SRGM bugs Markov Martingale total cost sub-optimal policy
Date of Defense	2007-03-29
Availability	unrestricted
Abstract Software testing is a process to detect faults in the completeness and quality of developed computer software. Testing is a key process in assuring quality by identifying defects in software, and possibly fixing them, before it is delivered to end-users. A major decision to make during this software testing is, to determine whether to continue testing and eventually releasing the software, or when to stop the test and ‘crash’ it. Such a decision needs to be made to optimally balance the tradeoff between the cost of development and the reliability of the software. In this paper, a new optimal strategy is developed based on a conditional non-homogeneous Poisson process (Conditional-NHPP) on a continuous time horizon to determine when the optimal time is to release or crash the software.
Files	Filename       Size       Approximate Download Time (Hours:Minutes:Seconds)     28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access    tanvirkhanthesis.pdf 716.86 Kb 00:03:19 00:01:42 00:01:29 00:00:44 00:00:03