
Term

Explanation

K




KepnerTregoe method

A structured decision aiding technique for collecting
information and making decisions developed by Charles H. Kepner and
Benjamin B. Tregoe in the 1960s. The approach consists of four basic steps,
situation appraisal, problem analysis, option analysis, and potential
problem and opportunity analysis.


key performance indicator (KPI)

A metric selected to indicate how effectively an organization is achieving its key business
objectives. The hundreds of KPI's that have been suggested may be
organized based on the type of organizational objective that they serve, for example, KPI's for financial objectives, customer
objectives, sales objectives, process objectives, staff
objectives, and so forth. To illustrate, examples of KPI's for staff objectives include
employee turnover rate, percentage of those who respond to open positions that are judged to be qualified for the position, and
employee satisfaction level. Organizations typically select and then use KPI's to help determine their progress toward achieving
strategic and operational goals, and also to compare performance relative to other organizations in the sanme industry.


knapsack problem

A mathematical statement of the problem of selecting
projects subject to a budget constraint. The name derives from the analogy
to the problem of choosing items to carry in a knapsack. To illustrate,
suppose there are m items, where item i has a
size c_{i} and, if selected, provides a benefit
b_{i}, for i = 1, 2,,...,
m. The capacity of the knapsack is C. The goal is to
select items that will collectively fit in the knapsack and provide the
greatest possible benefit. The problem may be expressed mathematically
as:
The x_{i} are decision variables (1
for item acceptance and 0 for rejection). These are the same equations used to describe the project portfolio
capital allocation
problem.
The knapsack problem is relatively difficult to solve. It
has, in fact, been used as the basis for encryption. Methods for solving
the knapsack problem are computationally intensive, however, various
approximate methods are available that are more efficient and that can be
shown to come very close to mathematically optimal solution (see Methods for Solving the Capital Allocation
Problem).
What makes the knapsack problem difficult is the "0/1
assumption"—Items must be put entirely in the knapsack or not
included at all. You cannot, for example, put part of a soda pop can in the
knapsack. Were it not for this requirement, you could solve the problem
easily using a "greedy algorithm"—Rank the items based on benefit per
unit size. Take as much of the topranked item as you can (or enough to
fill the knapsack). Then, repeat with the next ranked item until the
knapsack is full.
The knapsack problem arises in any activity, including
project portfolio management (PPM), that
requires allocating finite resources to items that are not infinitely
divisible. Thus, a relevant consideration for choosing a PPM tool is
whether it provides an algorithm for solving the knapsack problem and the
quality of that algorithm.

L




lean

As in lean project management, lean
enterprise, lean production, etc., a business practice that
considers the expenditure of resources for any goal other than the creation
of value for end customers (or in some uses, for the enterprise itself)
wasteful, and, therefore, a target for elimination. Toyota developed and
applied the concept to manufacturing in the 1990's, and the auto
manufacturer's success during this era helped popularize the practice.
Project portfolio management (ppm) applied
with the goal of selecting valuemaximizing project portfolios can be
regarded as consistent with the principles of lean operation, and some ppm
tool providers have used the phrase lean project portfolio
management to describe their offerings.


lexicographic method

A general approach to decision making or to prioritization
when multiple criteria are relevant. Specifically, the alternatives are evaluated relative to the criterion judged to be
most important, and the alternative with the best performance relative to this criterion is chosen (or ranked number one),
unless there are other alternatives that with regard to the most important criterion tie for first place. In that
case, evaluations with repsect to the second most important criterion are considered to break the tie. If
such comparisons don't resolve the tie, then the third most important criterion is consulted, and so on until one
alternative emerges as a winner. There are several variations on this approach which require more than the minimal
information of a strict lexicographic method. Note that contrary to most
multicriteria analysis
methods, lexicographic methods are not compensatory.


life cycle portfolio matrix

Also called the product life cycle portfolio
matrix and the ADL matrix, a simple tool developed in the 1980's
by the professional services firm Arthur D. Little intended to help a
company manage its collection of product businesses as a portfolio. The key
concept is consideration of where each product is within its business life
cycle.
Like other portfolio planning matrices, the ADL
matrix represents a company's various businesses in a 2dimensional matrix.
In this case, the columns of the matrix represents the growth stage of the
business product (embryonic, growth, mature, or aging) and the rows
represents the product's competitive position in the marketplace (dominant,
strong, favorable, tenable, or weak or nonviable). This results in a 4 by 5
matrix with 20 cells. The company's various product businesses are placed
within the matrix, and the positions are associated with logical business
strategies as shown below:
Life cycle portfolio planning matrix
The distribution and trajectory of the businesses across
the matrix helps indicate whether the firm's product mix is well balanced
now and in the future. For example, the company will need to maintain a
continuing set of mature businesses in order to generate cash to support
new embryonic and growth operations.


linear programming

A mathematical method for finding the maximum or minimum
solution to a problem where the objective function is a linear
combination of the decision variables, for example,
ax_{1} + bx_{2} +
cx_{3} ...
and where some or all of those variables are subject to
linear constraints, for example,
Ax_{1} + Bx_{2} +
Cx_{3} ≤ N or
Ax_{1} + Bx_{2} +
Cx_{3} ≥ N
Linear programming has been long applied to resource
allocation problems. Typically, the decision variables represent the amount
of various resources allocated to various purposes and the constraints
specify how much of each type of resource is available. So long as the
objective function is a linear function of the amounts allocated, the
solution can be found using linear programming.
The importance of linear programming derives in part from
the efficiency of the algorithm, known as the Simplex Method, by
which a linear program may be solved. Linear programming can handle very
large numbers of variables and constraints. Some applications, for example,
have involved millions of variables and hundreds of thousands of
constraints.
The main disadvantage of linear programming, of course,
is that it requires that the optimization be conducted based on a single,
linear objective function. Project
selection decisions, as well as most other decision problems, require
multiple, generally nonlinear, objectives to be simultaneously optimized.
(Goal programming and multiobjective linear programming are variations
of linear programming that attempt to account for multiple objectives.)
With linear programming you cannot, for example, account for project
startup costs, efficiencies of scale, and other considerations that
commonly cause the relationship between project costs and project benefits to be nonlinear.
Also, linear programming assumes that any solutions
within a continuum of possible values that satisfy the constraints are
possible (for contrast, see integer
programming). In the real world, most often you must choose whether to
fund or not fund a project. A linear programming solution that told you to
fund 1/5 of the project might not be very useful.


linear regression

The relation between variables in a regression analysis when the
regression equation is linear, e.g., y = ax +
b.


lottery

A term used on this website and by decision analysts and others to refer
to probability
distribution for some uncertain outcome, typically a monetary payoff.
For example, if the consequences of conducting a project are uncertain, the value of those consequences will likewise be
uncertain. A project selection decision model might be constructed to
derive a probability distribution over the values of possible project
outcomes. It might then be said that the decision to conduct the project is
like paying to play a lottery with outcome probabilities as specified by
the model.

