H. Michael Rauscher

Building Knowledge-Based Systems for Natural Resource Management

This is the first book to provide a detailed process for planning, designing, implementing, and testing knowledge-based systems for natural resource management. It presents material on all these major aspects of building a deliverable system. Equipped with these techniques, managers and scientists will improve their ability to solve complex resource problems that are multidisciplinary in scope and for which mathematical approaches prove insufficient. Fully describing the various components of these systems, this important work includes discussions on system design, knowledge acquisition, prototyping, knowledge verification and validation, implementation, and system delivery. To further illuminate the material presented, it contains a tutorial on the knowledge-based programming environment PROLOG as well as many examples of expert system development, including one for forest management. Building Knowledge-Based Systems for Natural Resource Management demonstrates how knowledge can be effectively organized and administered, enabling natural resource professionals to respond intelligently to natural resource problems. This book also provides researchers and students with an essential resource for understanding this useful technology.

Information and knowledge management in support of sustainable forestry: a review

Making good decisions can be extremely difficult when problems are not well structured and situations are complex, as they are when managing natural resources for multiple benefits and for users with differing values. Proficient problem solving depends on an adequate foundation of relevant and readily applicable knowledge. Recent advances in computer technology coupled with expanded knowledge and information distribution and accessibility, brought about by the Internet, have increased our power to manage both tacit and explicit knowledge. Impressive computer-based systems have been developed to deliver this knowledge for decision making, but their use in forestry has been limited. Widespread adoption will require close cooperation among people developing methodologies and techniques in the areas of inventory and monitoring, statistics and modeling, policymaking, and forest management planning. Without attention to the key task of knowledge management, however, efforts in sustainable forest management may only have limited long-term success. _______________________________________________________________________ Knowledge Management

SYLVATICA: an integrated framework for forest landscape simulation

Abstract In this paper we present the proposed conceptual approach and essential subsystems of SYLVATICA, a landscape-based simulation model which will integrate several resource-management technologies in a visual interactive environment. We believe this approach incorporates several critical technologies which should be common to any integrated landscape-based simulation and visualization environment.

Planning the Application

We have just completed the first section of the book. This included: motivation for the application of AI and KBSs to help solve resource management problems (chapter 1) and a look at the major components of knowledge-based systems (chapters 2 & 3). Now, in chapters 4–6 we turn our attention to the detailed steps of planning and designing a knowledge-based system, and acquiring and organizing the necessary knowledge to implement it. System implementation and evaluation issues are covered in chapters 7–10.

AI and Natural Resource Management

It has become increasingly common to hear or read news stories that involve the use or misuse of natural resources. Some examples are, for instance, the deforestation of tropical rain forests, timber harvesting and other activities on national forest (i.e. public) lands, or the loss of habitat for animal species. In most cases these stories include some controversy. Such disagreements arise for two primary reasons: (1) the complex character of natural resource phenomena and (2) the diversity of a resource’s user group. This book addresses the first of these, while the latter deals with social, political, and economic questions of extreme complexity that are only beginning to be discussed in academic as well as legal circles.

Designing the Application

In chapter 4, we discussed planning an application project and began with the problem definition. Red Pine Forest and Pest Management examples were introduced and defined. Once an application has been defined and knowledge sources identified, problem-solving knowledge must be acquired and organized. In the previous chapter, we reviewed some relevant knowledge acquisition concepts and methods to do this.

Other Knowledge System Components

In chapter 2, we introduced the important concepts of knowledge representation, reviewed inferencing methods grounded in symbolic logic, discussed methods to control the inference process, and introduced several knowledge system architectures. These issues make up the core of a knowledge-based system (Figure 3–1). To increase the utility of KBSs, we must also provide for explanations, interfaces both with the user and other computer software, and user-guided learning at run-time so that the domain knowledge available to the user can expand. These topics are discussed in this chapter. To round out our presentation of knowledge system components and techniques, we also review some ways to reasoning with uncertain knowledge. Following these topics, the remainder of the text examines how to apply those ideas to design, develop, and implement knowledge-based systems.

A PROLOG Toolkit Approach to Developing Forest Management Knowledge-Based Systems

In Chapter 8, native PROLOG was used to develop a complete knowledge-based system. Some of the implementation problems generated by the decision to limit ourselves to native PROLOG were discussed. In this chapter, we will look at how to use meta-PROLOG programming methods to build forest management KBS using a toolkit called DSSTOOLS. We will continue to use the Red Pine Forest Management Advisory System (RP-FMAS) introduced earlier as our implementation example, however, the focus is really on the toolkit. We have rewritten the original RP-FMAS using the PROLOG toolkit presented in this chapter. Readers unfamiliar with forest management theory and practice should review the background for RP-FMAS in section 4.3 and the knowledge model presented in section 6.1.1. Rauscher et al. (1990) and the RP-FMAS hypertext document (see the Frontmatter for file location) may be consulted for additional detail.

Knowledge-Based Systems: Representation and Search

In Chapter 1, we briefly introduced the idea of knowledge-based systems and included a simple structural representation for such a system (Figure 1–6). This chapter and the next will focus more closely on those ideas and introduce new ones using the more comprehensive knowledge-based system architecture in Figure 2–1. Examples of specific implementations of these concepts (using the Prolog language) can be found in chapters 7, 8 and 9.

Programming Knowledge Systems in PROLOG

After planning an application and developing the problem definition in chapter 4, acquiring the domain knowledge from sources in chapter 5, designing the knowledge model and human factors model in chapter 6, and selecting PROLOG as our major AI development tool, we need to pause before continuing to develop prototype systems in chapter 8 and 9 (Figure 8–1). In the present chapter, we discuss the strengths and weaknesses of PROLOG and provide an introduction of the main elements of the PROLOG language. If the reader has never programmed in any language before, we suggest this chapter be skipped because it will convey little useful information. On the other hand, readers who are programmers in other languages, should be able to understand the essence of PROLOG from this introduction. Those already familiar with PROLOG programming may regard this chapter as a useful review.