Data Perturbation Analysis for IS Project Management Based on a Single Time Estimate

The Critical Path Method (CPM) is the most widely used tool for project management; however, it requires three estimates for the duration of each activity as its input. This is too much uncertain input requirement from the managers, making the critical path (CP) unstable causing major difficulties for the manager. A linear programming formulation of the project network is proposed for determining a CP, based on making one estimate for the duration of each activity. Upon finding the CP, data perturbation analysis (DPA) is performed using the constraints of the dual problem. This DPA set of uncertainties provides the manager with a tool to deal with the simultaneous, independent, or dependent changes of the input estimates that will preserve the current CP. The proposed procedure is easy to understand, easy to implement, and provides useful information for the manager. A numerical example illustrates the process. DOI: 10.4018/jeis.2012100104 78 International Journal of Enterprise Information Systems, 8(4), 77-98, October-December 2012 Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. runaway information system project examples. In 1982, a major insurance company began development of an $8 million computer system from a major software provider. This system was intended to serve all of the computing needs of the insurance company and was due to be completed in 1987. However, a number of problems were encountered resulting in delays of completion until 1993, with a new estimated cost of $100 million. Pirdashti, Mohammadi, Rahimpour, and Kennedy (2008) used Delphi technique to elicit expert opinions about criteria for evaluating the network locations for a network design project. Projects are systems consisting of interrelated parts working together to accomplish project objectives. There are a number of important roles within information systems projects. Project managers have to balance technical understanding with the ability to motivate diverse groups of people (the project team) brought together on a temporary basis. Projects are collections of activities. If one activity is late, other activities are delayed. If an activity is ahead of schedule, workers tend to slow down to meet the original completion date. Information systems projects have many similarities to generic projects. They consist of activities, each with durations, predecessor relationships, and resource requirements. They involve high levels of uncertainty and often suffer from time and cost overruns, while rarely experiencing time and cost under runs. However, information systems projects are different from generic projects in some aspects. While each project is unique, there are usually many, many replications of information systems project types. Most are served by a standard methodology, with the need to identify user requirements, followed by design of a system, production of the system, testing of the system, training and implementation, and, ultimately, maintenance of the system. These steps are not always serial; there are often many loops back to prior stages. Defining project success is in itself difficult. There are many views of what makes a project successful. Successful implementation has been found to require mastery of the technical aspects of systems, along with understanding key organizational and behavioral dynamics. There has been a great deal of research into information systems project failure. Failure can occur when design objectives are not met. The difference between successful and failed information systems projects often lies in planning and implementation. A great deal of research has been performed to identify factors that lead to project success. These factors include planning, user involvement, and good communication. Additional factors that are reported as important in information systems project success repeatedly include top management support and clear statement of project objectives. Information systems project management can involve a wide variety of tasks. Typical information systems project types include maintenance work, conversion projects, and new systems implementation. Maintenance projects are by far the most common type of information systems project. They can arise from need to fix errors or to add enhancements to some system, or they can involve major enhancements. Conversion projects involve changing an existing system. Most conversion projects consist of moving an application from one computer platform to another. New systems development involves different management characteristics by type of system. The types of systems include transaction processing, management control, decision support systems, data warehousing and data mining, group support system, executive information systems, Internet commerce, and enterprise resource planning (ERP). Ahmad, Zakaria, and Sedera (2011) state that companies face the challenges of expanding their markets, improving products, services and processes, and exploiting intellectual capital in a dynamic network. Therefore, more companies are turning to an enterprise system for support across the entire lifecycle, from selection and implementation to use. Chen and Wang (2012) are aware of numerous difficulties associated with learning in a project. For facilitating knowledge learning, an integrated project management model is constructed. Hanafizadeh, Gholami, International Journal of Enterprise Information Systems, 8(4), 77-98, October-December 2012 79 Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Dadbin, and Standage (2010) also considered the implementation of enterprise resource planning systems, requiring huge investments, while ineffective implementations of such projects are commonly observed. A considerable number of these projects have been reported to fail or take longer than initially planned, while previous studies show that the aim of rapid implementation of such projects has not been successful; and the failure of the fundamental goals in these projects have imposed huge amounts of costs on investors. Sternad, Bobek, Dezelak, and Lampret (2009) are aware of the facts that ERP implementation is a complex process that requires substantial resources and efforts, and yet the results are very uncertain. Lai (2006) rightly emphasizes that it is very important to identify what are the key factors across different steps within the ERP implementation. Zarei and Naeli (2010) give warning that it may entail new hazardous challenges if enterprise resource planning cannot be well managed to achieve success. Critical success factors are project management, top management support, business process reengineering, change management, and training. Orlowski and Kowalczuk (2006) explicitly outline the difficulties involved in software project management, in particular regarding the planning and control of processes and the project-team manager. Agile methodology is an approach to project management software development for dealing with business realities, such as changing requirements during development (Kendall, Kong, & Kendall, 2010). Information systems projects have high levels of uncertainty. The size of the project is usually not well understood until systems analysis has been completed. Most of the unexpected delay in these projects occurs during the latter stages of testing. Almost one third of the time used in typical projects was required for the planning phase, and coding typically consisted of one-sixth of the project. Coding is the most predictable portion of the project. The last half of the project was testing – onequarter for component testing and one-quarter for system testing. The activity most difficult to predict was testing. Project managers currently use the critical path method (CPM) and/or project evaluation and review technique (PERT) in order to provide a planning and control structure for projects. CPM helps managers understand the relationships among project activities (see a sample of general project management textbooks and those with MIS applications included in the References section). Key personnel and other resources can be allocated to activities on the critical path. These activities can be closely monitored to avoid completion delays. CPM identifies the sequence of activities that will have the longest completion time of the entire project. PERT extends CPM’s scope to deal with uncertainties inherent in any project. Project activity network models ranked above all other quantitative decision making tools in terms of the percentage of firms who used them. Although these techniques are widely used in practice and have provided economic benefit to their users, they are not problemfree. Problems derive from the nature of the underlying statistical assumptions, the potential for high computational expense, and from the assumption that factors determining activity duration are essentially probabilistic rather than determined by managerial action. Critique of CPM/PERT Here “critique” means the scope and limitations. The most widely used CPM/PERT algorithm is based on the following assumptions and procedure: • Activity duration is rarely (if ever) known with certainty; the CPM asks the manager for three time estimates; • Three time estimates are determined (by guesses, etc.) for each activity; ◦ a = an optimistic completion time; ◦ m = a most likely completion time (this is what we will use in this paper); ◦ b = a pessimistic completion time; • Activity approximations; 80 International Journal of Enterprise Information Systems, 8(4), 77-98, October-December 2012 Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. ◦ An approximation for the distribution of an activity’s completion time is a