Lee Merkhofer Consulting Priority Systems

Technical Terms Used in Project Portfolio Management (Continued)






























Stands for generally accepted accounting principles—a collection of rules, procedures and conventions, established by various government and private organizations, to ensure that a company's financial statements accurately convey its financial status. Some vendors of project portfolio management tools advertise that they incorporate metrics (or have other features) that facilitate compliance with GAAP.

game theory

A branch of applied mathematics that analyzes the behavior of individuals who are pursuing self interest against other individuals who are doing the same. Game theory addresses situations in which an individual or organization chooses actions in situations where the consequences of those actions depend on the choices made by others. Applications of game theory typically attempt to find an optimal strategy for one player to use when an opponent is also assumed to be playing optimally. Such strategies are termed equilibrium strategies in that they are stable and unlikely to change. Game theory is used to support decision making in many different areas, including business and military strategy. Some project portfolio management tool intended for application to investments in highly competitive environments utilize consequence models based on game theory.


In project prioritization, the phenomenon wherein individuals deliberately bias estimates in order to improve the evaluation of favored projects.

Gantt chart

A horizontal bar chart that displays the timing, duration, and interactions among multiple, time-phased activities, tasks, or projects. Gantt Charts, named after the originator Henry L Gantt, are a useful project management and planning devise.

Sample Gantt chart

A simple Gantt chart

As shown in the example, a Gantt chart is a essentially a table with each row corresponding to an activity. Time (e.g., measured in days, weeks, or months) is denoted by the columns. Each task is represented by a bar extending across the time columns, indicating the planned duration of the task. Milestones and critical path lines are used to add further detail to the chart. Milestones are important checkpoints or deadlines and are indicated by small symbols in the time columns. Critical path lines connect task bars to indicate dependencies, such as the requirement that a task be initiated after the completion or commencement of another task.

Gantt charting capability is routinely provided in tools for project management and in many project portfolio management tools.

gap analysis

A systematic comparison of the current situation to the desired state. Gap analysis generally includes benchmarking and other assessments for clarifying expectations. The end goal of gap analysis is the development of specific plan for closing the gap and moving organizational performance to the desired state.

genetic algorithm

A mathematical search technique for solving optimization problems based on an algorithm consisting of steps similar to those that occur in natural evolution. The technique has found applications in numerous fields, including biology, engineering, economics, and manufacturing.

Basically, a genetic algorithm seeks an optimal solution by simulating an evolutionary process. An initial "population" of potential solutions is randomly generated. Each candidate solution in this initial "generation" is evaluated to determine its "fitness." Those determined most fit "survive" and are combined and mutated to form a new population (the next generation). The new population is then similarly used to conduct the next iteration of the algorithm. The process continues until either a maximum number of generations has been produced or an acceptable fitness level has been obtained for a solution.

Basic genetic algorithm

Flowchart of the basic genetic algorithm

Optimization based on generic algorithms is applicable to project portfolio management (ppm), and some ppm tools use the technique. In this context, an initial generation of potential portfolios is randomly generated (parent portfolios). Portfolios that don't meet the constraints (e.g., cost constraints) are eliminated (they die off). Pairs of individual portfolios are then combined (e.g., by choosing every other project from each) to produce second-generation portfolios (child portfolios). Child portfolios that don't meet the constraints are eliminated. The remaining portfolios are then evaluated, and the highest ranked are selected to represent the next generation of portfolios. This process is repeated for a set number of iterations or until the user-specified optimization parameters are satisfied.

A common problem with optimization algorithms is premature convergence—the optimizer homes in on a solution that is not really optimal. This occurs with genetic algorithms because the population of potential solutions being used loses diversity. To address this problem, an approach is used that mimics the way nature maintains diversity. New portfolios ("genetic mutations") are randomly generated and periodically introduced into the portfolio populations. The mutated portfolios only "survive" if they meet the constraints, in which case they help keep the genetic algorithm from converging to a false optimum.


A desired result or end to which effort is directed. The words objectives and goals are sometimes mistakenly used used interchangeably. However, in technical usage the word objective typically refers to a precise, well-defined desire (expressed so as to indicate an object of value, context, and the direction of preference), whereas the word goal may refer to a longer-term, less concrete aim. Objectives are measurable, whereas goals may not be. For example, learning more about Chinese history might be a goal whereas obtaining a high score on the Chinese history test might be an objective. In other contexts, the word goal is used to specify a target level of achievement against an objective, for example, "obtain a score of at least 95 on the history test."

goal factoring

An individual or organization will typically have multiple goals. Goal factoring is the process of structuring goals, which can aid the identification, evaluation, and prioritization of possible actions for achieving those goals. One first lists goals, then organizes them in relation to each other, identifying the goals that are most desired and those that might be complementary or conflicting. Typically, goals are arranged graphically in a diagram with lines or arrows indicating relationships.

Example goal factoring

Example output of goal factoring

Goal factoring is basically a less-rigorous version of the methods for modeling decision-maker preferences used by more formal decision-aiding methodologies (such as multi-attribute utility analysis and AHP). For example, goal factoring diagrams can appear similar to objectives hierarchies, although the goals appearing in the diagrams typically do not meet all of the requirements (such as being measurable, non-overlapping, etc.) necessary to allow the diagrams to be converted to quantitative decision models. Even so, in some contexts (e.g., financial investment portfolios), the term goal factoring has been used to refer to a process of deriving a single metric that accounts for portfolio risk, timing, return, and other relevant measures of portfolio performance.

goal programming (GP)

An optimization method designed for problems with more than one objective and discrete or continuous decision variables that allows a solution to be found using linear programming. The method, sometimes applied to resource allocation and project selection problems, involves expressing goals or targets for objectives and assigning priorities or weights to achieving those targets. For example, the formulation might be to find the subset of R&D project opportunities that comes closest to achieving specified goals for manpower utilization, market share, sales, and net present value maximization. Constraints on acceptable solutions may also be defined, for example, requiring total costs to be no greater than the budget. Goal programming (GP) seeks solutions that meet constraints while minimizing the weighted sum of the deviations from the specified targets. The solution effectively involves a repetitive process of attempting to achieve each goal, in order of priority, subject to the specified constraints.

GP has several attractive features. One obviously, is the ability to address multiple objectives. Also, and importantly, solutions can be easily found using the Simplex Method of linear programming. This means that relatively large numbers of decision variables, constraints, and goals may be established without creating difficulties for finding a solution. For example, in the context of resource leveling, you could solve for multi-year solutions that closely achieve many detailed goals with many specified constraints.

A feature that no doubt promotes the use of GP is the (apparent) non-demanding nature of the necessary inputs. Like other prioritization logics, a model relating choices to performance is required. However, GP does not require detailed quantification of decision making preferences and willingness to make tradeoffs the way that most multi-criteria analysis methods do. Instead, all that is required is the specification of goal targets and weights.

The main disadvantage of GP, of course, is that reasonable goals and targets cannot be specified without reference to underlying decision-maker preferences—choosing the "right" targets and weights is exceedingly difficult. If the targets and weights are not appropriate, the solution will not be the one in the best interest of the organization. The reason for this is that GP programming does not allow tradeoffs between goals. For example, if sales growth is the first priority goal, and market share is the second, the formulation implies that not even one dollar of sales growth can be sacrificed to obtain even a huge gain in market share. GP represents a "satisficing" approach to decision making, meaning that what is sought is a satisfactory solution rather than one that is truly optimal. Because the method does not capture willingness to tradeoff achievement of the various objectives, it is incapable of finding solutions that lie on the efficient frontier.

Goal programming has seen applications in production planning, scheduling, health care, portfolio selection, distribution system design, energy planning, water reservoir management, timber harvest scheduling, and wildlife management problems. Many of these applications have been used in combination with other methods to accommodate the proper weighting of criteria.


The means by which an organization regulates and controls organizational behavior in accordance with its goals and objectives. A governance structure establishes accountability by implementing systems to monitor and record what people do, includes steps for ensuring compliance with policies, and provides for corrective action in cases where rules have not been appropriately followed.

grid analysis

A decision-aiding technique that involves creating a table with options listed as rows and factors that need to be considered as columns. Each option/factor combination is then scored, the scores are weighted, and the results added to provide an overall score for the option. The approach, although simplistic, is used by some project portfolio management tools to rank projects.

group think

A dynamic of groups that promotes faulty decision making. Group think, a term coined by psychologist Irving Janis in the 1970s, has been widely studied. Consequences of group think include the tendency of groups to overlook alternatives, selectively collect information, fail to anticipate adverse consequences of choices, assume agreement among members when it does not exist, and fail to develop contingency plans. In addition, studies show that groups tend to possess a sense of invulnerability which promotes risk taking, and causes members to stereotype the views of those outside the group, self-censor their own views if they go against the group view, and believe in their inherent moral superiority over those outside the group.


Pronounced GOO-ee, GUI stands for graphic user interface and refers to the common method used for enabling humans to interact with a computer program. A GUI is graphic-based and obtains user inputs at least in part via icons, pictures, and menus (which the user designates with a mouse or other pointing device) as well as through keyboard-entered text.