The Analytic Hierarchy Process

Research and development (R&D) is often a well spring of new ideas and concepts leading to the development of commercially viable products and processes. An organization’s future market and financial positions may depend in large measure on the R&D project proposals which are selected. A variety of financial, market, technical, and manufacturing criteria may influence the selection decision. The importance of specific criteria varies by type of R&D activity, and the extent to which a particular project supports business objectives. This paper describes an approach for modeling the R&D project selection decision using the Analytic Hierarchy Process. The AHP represents an improvement over other well-known scoring approaches since the criteria weights or priorities established by the AΗΡ are not based on arbitrary scales, but use a ratio scale for human judgments. The paper begins with a brief review of the R&D project selection literature, leading to a description of the desired characteristics for a decision support system for project selection. For a specific R&D strategy, namely, new product development, an AHP model is developed using an illustrative example. For situations requiring a large number of projects to be evaluated, the AΗΡ model is expanded to include a series of performance ratings for each criterion. The performance ratings and weights for each criterion are transferred to a spreadsheet program which produces the final project rankings. The resulting project priorities or scores are included in an integer programming model to assist in the project funding decisions. The relationship between the integer programming approach and a form of benefit-cost analysis is discussed and illustrated. Two extensions of the AHP approach are then presented. The first addresses situations requiring the evaluation of a broader set of project selection criteria. The second illustrates how the AHP project selection model can be linked to the strategic planning process through an analysis of the mission, objectives, and strategies of the business. The paper concludes with a discussion of future areas of research.

The use of the Analytic Hierarchy Process as part of a decision aid for research and development (R&D) project selection in a large Finnish company is described. The company’s most important operational sector is the oil industry and a comprehensive petrochemical industry has been built as an extension to oil refining. Selection of R&D projects is a decision problem of vital importance in the company’s long-range strategy. To evaluate one project proposal at a time, a subjective measurement scale is constructed for each lowest level criterion. The importance of the selection criteria as well as the measurement scales are assessed by the AΗΡ or direct rating. The overall preference model is implemented in a decision aid by integrating different software modules. The preference models of the individual managers reside in a database which is managed by a database program. These models can be updated by a general purpose decision analysis program. The man-machine interface of the system is implemented by an expert system shell.

This paper describes a method for identifying research needs and obtaining priorities for them. The overall method consists of two parts. The first is concerned purely with the identification of research needs without regard to their relative importances or urgencies. The second part is devoted to obtaining priorities for the identified needs via the Analytic Hierarchy Process. A case study is presented in which this method was used to develop a Master Plan for Research in surface hydrology and surface water resources for South Africa.

The Analytic Hierarchy Process can be used to design rates to fulfill specific goals for ratemaking. To achieve these goals, a costing methodology must be chosen among various available types including those based upon accounting costs and marginal costs. The AHP can be employed with companies, regulators, and customers to reconcile conflicting interests in making these choices among costing methodologies. Since construction costs have a major impact upon future rates, accurate load forecasts are essential for adequate ratemaking. Again, the AΗΡ presents a simple and rational decision aid for making such forecasts. In the cases of ratemaking and load forecasting, the AHP is much easier than traditional methods that regulators have used to make these decisions.

Over the last 20 years, the debate among legislators, regulatory officials, manufacturing associations, and environmental groups over an acid rain policy for the United States has yet to be resolved. The high level of disagreement among the knowledgeable participants about the complex nature of the acid rain problem and about appropriate solutions has stalled repeated attempts by the U.S. government to reach a consensus of opinion on which policy option is best. Yet, the outcome of each year’s debate and the resulting policy can dramatically impact the operations of electric power companies and manufacturers of aluminum, steel, and automobiles. Faced with the current impasse, these firms still need to predict which acid rain policy might be adopted so that they can formulate an effective yearly business plan.

In this paper, we present an AHP-based model that is designed to help power company decision makers predict a national acid rain policy. We envision our model and the accompanying analysis as forming a decision support template that can be regularly updated by management to gain new insights about acid rain policies under consideration. The results of this modeling exercise can be used by power companies to help plan capital budgeting decisions related to the timing and design of new electric power plants, as well as to specify required “clean air” modifications to plants already in operation.

This paper discusses a decision support tool that can help players of a war game simulation decide on the most efficient use of military forces and limited airlift and sealift required to move the military forces. The Analytic Hierarchy Process is employed to develop military unit values that are used as coefficients in a linear program that assigns forces to different regions of the world during crisis decision exercises.

The AHP has been applied to the highly complex problem of determining the adequacy of security at selected U.S. Department of Energy facilities that produce nuclear weapons. Use of the AHP enabled an Inspector General’s team to determine the relative importance of specific countermeasures at these facilities. Thus, the team was able to weight and rank recommendations for corrective action where deficiencies were found to exist. The hierarchy, which included a variety of factors, provided the Department of Energy with a flexible tool for assessment, planning, and allocation of resources for enhanced security.

Over the last eight years, the Canadian management consulting firm of Woods Gordon has used the Analytic Hierarchy Process to help clients in the public and private sectors structure and solve complex, real-world decision problems. Many of the applications were costly projects that required detailed hierarchies with a large number of criteria and alternatives and that often involved a group of decision makers.

In this paper, we review the use of the AΗΡ in a consulting environment and focus on the many practical considerations that users must address in order to facilitate a successful decisionmaking process. To provide some background on how the AHP works in practice, four applications are described in detail: a hospital’s building and renovation program, strategic planning for information systems needs, contractor selection, and the allocation of military maintenance work.

An AHP model was used to develop antibiotic treatment guidelines for young women hospitalized for acute pyelonephritis. Seven antibiotic regimens representative of current treatment recommendations were compared relative to four major criteria. The resulting analysis identified a combined regimen of ampicillin and gentamicin as the best choice for initial treatment pending results of urine culture and antibiotic sensitivity testing. The use of this regimen was recommended to a group of physicians. Subsequently the use of ampicillin and gentamicin increased significantly in young women with pyelonephritis. This study shows that a significant change in the process of patient care was associated with treatment recommendations based on the AHP. This finding indicates that the AΗΡ may be a valuable tool for helping physicians make better, more logically consistent patient management decisions.

In this study, we model the Northern Ireland conflict using the Analytic Hierarchy Process. Our model is an update of earlier analyses carried out in 1976, 1977, and 1982 in which it was shown that the outcome which would most satisfy the aspirations of all parties would be legislative independence for Northern Ireland. The current analysis takes into account important changes that have taken place since the earlier work was performed. We show that the most satisfactory outcome is still one of legislative independence.

A Japanese manufacturer of electronic goods is planning to construct a new, state-of-the-art factory to manufacture large scale integrated circuits. A team of company managers must decide on the appropriate location for the factory taking many different criteria into account. This paper reports on the decision process that the project team employed to model the site selection problem.

This paper provides a summary of a project done in 1986 for an organization in a Central American country. The purpose of the work was to develop a global strategy and some functional strategies to improve the organization’s image and to generate new sources of funds.