Probabilistic estimation of software project duration
📝 Abstract
This paper presents a framework for the representation of uncertainty in the estimates for software design projects for use throughout the entire project lifecycle. The framework is flexible in order to accommodate uncertainty in the project and utilises Monte Carlo simulation to compute the propagation of uncertainty in effort estimates towards the total project uncertainty and therefore gives a project manager the means to make informed decisions throughout the project life. The framework also provides a mechanism for accumulating project knowledge through the use of a historical database, allowing effort estimates to be informed by, or indeed based upon, the outcome of previous projects. Initial results using simulated data are presented and avenues for further work are discussed.
💡 Analysis
This paper presents a framework for the representation of uncertainty in the estimates for software design projects for use throughout the entire project lifecycle. The framework is flexible in order to accommodate uncertainty in the project and utilises Monte Carlo simulation to compute the propagation of uncertainty in effort estimates towards the total project uncertainty and therefore gives a project manager the means to make informed decisions throughout the project life. The framework also provides a mechanism for accumulating project knowledge through the use of a historical database, allowing effort estimates to be informed by, or indeed based upon, the outcome of previous projects. Initial results using simulated data are presented and avenues for further work are discussed.
📄 Content
CITATION: Connor, A.M. (2007) “Probabilistic estimation of software project duration”,
New Zealand Journal of Applied Computing & Information Technology, 11(1), 11-22
Probabilistic Estimation of Software Project Duration
A.M. Connor Auckland University of Technology andrew.connor@aut.ac.nz
Abstract This paper presents a framework for the representation of uncertainty in the estimates for software design projects for use throughout the entire project lifecycle. The framework is flexible in order to accommodate uncertainty in the project and utilises Monte Carlo simulation to compute the propagation of uncertainty in effort estimates towards the total project uncertainty and therefore gives a project manager the means to make informed decisions throughout the project life. The framework also provides a mechanism for accumulating project knowledge through the use of a historical database, allowing effort estimates to be informed by, or indeed based upon, the outcome of previous projects. Initial results using simulated data are presented and avenues for further work are discussed. Introduction Estimation of cost and duration for software development activities is one of the most difficult aspects of software project management. The project manager often has the responsibility to make accurate estimations of effort and cost against which a project’s success will be judged. This is particularly true for projects subject to competitive bidding where a high bid could result in losing the contract or a low bid could result in a loss to the organisation. From an estimate, the management often decides whether to proceed with the project. Industry has a need for accurate estimates of effort and size at a very early stage in a project.
This paper, which extends an earlier conference paper (Connor & MacDonell, 2006), outlines a methodology for introducing probabilistic modelling for the estimation of duration for software development projects. Software development, more so than many other disciplines, is plagued by vague or shifting requirements and a lack of understanding regarding product complexity that often leads to projects being delivered either late, over budget or not to requirements. Software cost estimates made early in the software development process are often based on wrong or incomplete requirements.
In this paper, uncertainty in effort estimates are linked to a project work breakdown
structure in an effort to achieve two purposes. Initially, the method described in this paper
can be utilised during the development of a tender submission for a software project.
Typically in this circumstance, a project effort estimate will be made using a number of
methods, such as expert opinion or a parametric model. It is possible that some
companies will base a bid/no-bid decision on this single estimate without any deeper
analysis of the risks involved. The tool detailed in this paper, developed in Excel using a
freely available add-in, SimulAr (Machain, 2005), allows uncertainty in estimates to be
captured and use of Monte-Carlo simulation provides an indication of the range of likely
outcomes, not just a single estimate. The bid/no-bid decision can therefore be informed
CITATION: Connor, A.M. (2007) “Probabilistic estimation of software project duration”,
New Zealand Journal of Applied Computing & Information Technology, 11(1), 11-22
by pessimistic, optimistic and realistic estimates. In addition, the analysis of the data
allows the areas of highest risk to be located and as such the project manager can allocate
resource in the development of the tender submission in order to reduce this risk to an
acceptable level.
Following a successful tender submission, the same process can be used to track and refine cost information to track progress and continue to highlight the potential risk areas. Subject to the constraints of the development process itself, it is feasible to re-allocate resource and re-order tasks in the process to reduce the risk in certain areas to bring a wayward project back on track. Larger organisations that have multiple projects running simultaneously can utilise the information provided by the tool to manage the risk and resource across their portfolio of projects.
A key feature of the tool is its ability to capture and utilise project duration data for use in providing more accurate estimates for future projects. The use of such corporate knowledge is particularly appropriate for organisations that produce variants of a product or undertake very similar projects. The number of organisations that may fully utilise this feature will depend greatly on the environment, and it may be most applicable for larger companies outside of New Zealand. To address this, the tool does not mandate the use of historical data therefore allowing it to be applied to both typical and atypical projects. For atypical projects, the underlying
This content is AI-processed based on ArXiv data.