chapter 4 distributive regulation frameworkeil.stanford.edu/publications/shawn_kerrigan/temp... ·...

Chapter 4 Distributive Regulation Framework

32

Chapter 4 Distributive Regulation Framework

4.1 Introduction

Hazardous waste regulation compliance checking is a distributed process conducted among regulators (EPAs), waste generators (the units that generate waste), and waste Treatment, Storage and Disposal Facilities (TSDFs). During the hazardous waste compliance checking process, individuals from these parties with different professional backgrounds and domain expertise must collaborate with one another and make the decisions about the regulation compliance. Each participant is responsible for particular tasks that are of importance to the compliance checking process. Currently, the hazardous waste regulation compliance checking is performed manually, with exchange of information that is recorded in written forms.

In order to understand the compliance process, we conducted an analysis of the information flow among the participants. We use one of the compliance checking tasks, namely, identification for hazardous waste for waste generator to illustrate the details. The process in identifying waste for a generator depends on regulations from different sources; the regulations are from federal, state and, in some states, local regulators. Because of the complexities involved in applying the regulation codes for compliance checking, generators usually contract TSDFs to handle waste identification and other related waste management tasks. TSDFs specialize in waste management and provide procedures for dealing with the waste identification and compliance assistance for the generators. The process can be depicted as shown in Figure 4.1.

To comply with regulatory requirements related to hazardous wastes, the generator follows the regulation codes from the EPAs and the interpretation of the codes from the TSDFs. There is an active interaction and exchange of information and knowledge among the three parties: the waste generator, TSDF and EPA. This chapter attempts to formalize the information flows and the interaction among the participants. Specifically, issues are studied from three perspectives: (1) identifying the information and knowledge in the waste regulation compliance process, (2) developing a methodology suitable for the waste regulation compliance process and (3) establishing a distributed computing framework for waste regulation compliance process.

Regulator Functionality

Enforcement of of Regulation codes

Facilitator (TSDF) Functionality

Application of Regulation codes

Obtaining a generator’s waste information and checking

compliance on behalf of a generator.

Generator Functionality

Compliance with Regulation codes

Information and knowledge flow

Information and knowledge flow

Figure 4.1: Information flow during hazardous waste regulation compliance process

4.2 Code Compliance from Different Perspectives

The regulation codes are the basis for the waste generators to conduct waste identification and to perform hazardous waste compliance. Ironically, a waste generator is most likely the one that has the least understanding of the regulation codes. Furthermore, regulation codes change frequently and it is unlikely that a generator is updated for the changes and uses the latest codes for checking. The regulators, say federal and state EPAs, are the original drafters of the codes and the written codes, with their best intention, often do not provide the generators a clear interpretation of the code. In fact, it is not feasible for the regulators to provide comprehensive instructions how to use the codes and to develop compliance procedure. The difficulty lies in the different levels of understanding and appropriate interpretation of the regulation codes. The situation can be illustrated by examples for a waste generator to identify the hazardous waste that it produces.

The identification of a hazardous waste for a waste generator should follow the requirement of the Code of Federal Regulations (CFR) part 261, 262. (40 CFR Part 261 and Part 262). The provisions in these parts are written in legal language that contains vagueness and incomplete information. For example, the second provision of Section 40 CFR 262.11 (c) reads:

Section 262.11 (c) (2) Applying knowledge of the hazardous characteristic of the waste in light of the materials or the processes used.

Here, the phrase " knowledge of the hazard characteristic of the waste" is vague, so is the phrase "in light of the materials or the processes used." These phrases can be interpreted in many ways, and therefore, the meaning of this provision is basically undetermined. Furthermore, the information contained in the provision is incomplete and it does not provide any references on where and how to find further information to clarify the vague phrases. If a generator tries to use these provisions to characterize a hazardous waste, the generator needs further interpretation of the provisions.

From the interviews with several waste generators, which include Hewlett-Packard, Intel, and Text Instrument, in the semiconductor manufacturing industry, most generators consult their TSDFs whenever there is question whether a certain waste is hazardous. The legal professionals in TSDFs specialize in the hazardous waste regulations and provide guidance for a waste generator to identify a waste or wastes. From the interviews with a TSDF, Romic Environmental Technologies, who works with waste generators in the semiconductor industry, when a TSDF receives a request from a generator for identifying wastes, the legal professionals in the TSDF usually search their knowledge base for the hazardous waste identification information and the related regulations. The information, knowledge interchange and collaboration are conducted manually.

The waste identification process can become quite complicated because the state regulations could be different from the federal regulations. For example, the California Code of Regulation is stricter than CFR in that there are wastes that are deemed hazardous in California but are not considered hazardous by the federal government. In California, the regulator, namely, the Department of Toxic Substances Control of California EPA, uses a Chemical Cross Index, called List of Lists, to link the waste and the related chemical components to a certain regulation and then uses the regulation to identify whether a waste is hazardous.

As discussed above, different participants have different knowledge about the regulation codes and follow different procedures in the regulation compliance process. To make regulation compliance practical, most generators rely on their TSDFs for a complete interpretation of the regulation codes that are applicable to them. In this sense, TSDFs serve as mediators between the generators and the regulators. The assumption here is that the TSDFs have good knowledge of the regulation code and their interpretations are accurate. Furthermore, it is also assumed that TSDFs have sufficient knowledge about the applicability of the regulation codes.

This information flow among regulators, generators and TSDFs is shown in Figure 4.2. During the compliance checking, a TSDF works as a mediator for assisting generators to comply with regulation codes. At the same time, a TSDF usually communicates with regulators on code revisions and new regulatory requirement. The acquired information is kept as local knowledge base. When there is a waste identification request, a TSDF can then use its local knowledge source to provide appropriate information to generators.

TSDF

Regulator

Generator

1. Request for waste

identification

2. Ask for the latest

regulation codes

3. Publish the latest

regulation codes

4. Provide the

information of how

to use the related

codes and to check if

the waste is

hazardous in terms

that a generator can

understand

Figure 4.2: Information and knowledge flow in identifying a generator's waste

In this chapter, we focus on building a distributed information management paradigm to enable information flow and interaction among the different parties during hazardous waste compliance checking. There are several approaches to deal with complex information flow and interactions. One approach is to employ domain based information representation [Russell and Norvig, 1995] to express the information exchange interfaces and the information representation within a domain. However domain based representation does not support the interaction among diverse information sources in a distributed information paradigm. Another approach is to employ context, hierarchy of contexts, and information and knowledge interchanges among contexts to describe the interaction among the different information sources [McCarthy, 1994; McCarthy and Buvac, 1997]. In the context-based representation, it is argued that in a complex information and knowledge organization and environment, the participants organize and use knowledge for their own purpose in their individual subcontexts. The information and knowledge created from individual subcontexts does not support direct sharing in a more general context [McCarthy, 1994]. Therefore, we need to make explicit the individual context when sharing information among multiple participants.

As noted, in the hazardous waste compliance process, the information and knowledge for each participant has its own local resources that could be different from one another. In this research, we employ a context-based framework for describing the organization, interaction, and integration of the distributed information sources to support the hazardous waste regulation compliance process. The purpose for using contest based information and knowledge organization is threefold: (1) to formally formulate the distributed information and knowledge interchange during the compliance process, (2) to formally distinguish knowledge in the compliance process and the background knowledge for understanding the information being used, and (3) to present the information infrastructure for compliance checking process from an engineering perspective. Specifically, we introduce an Context Based Information and Knowledge Organization (CBIKO) for formalizing the distributed information and knowledge framework for the hazardous waste regulation compliance checking.

4.3 A Context-Based Model for Distributed Regulatory Information

In this work, a context-based concept is employed to organize the distributed information and knowledge for the hazardous waste compliance process. Contexts are organized locally from the perspective of an individual party and globally to coordinate the information and knowledge sources among all parties.

In Sections 4.3.1 and 4.3.2, we discuss the issues for forming the local context of regulation compliance checking for participants. In Sections 4.3.3 and 4.3.4, we discuss the issues for forming a global context for regulation compliance checking. A mediation [Wiederhold, 1994] approach is then proposed to facilitate regulation compliance checking process, using the context based organization of regulation resources.

4.3.1 Information and Knowledge Organization in Compliance Process

4.3.1.1 The Compliance Knowledge of a Participant

During the hazardous waste regulation compliance checking process, different participants have different knowledge and very often, different understanding about the regulation codes, the checking procedure, and the related documents. The differences can be illustrated using the regulation codes for identifying hazardous waste. A general process of waste identification for a generator can be summarized as shown the Figure 4.3. When identifying a waste, a generator encounters environmental law in the form of statutes and regulation codes [Kindschy, kraft and Carpenter, 1997]. The first step is to determine which regulation codes to use, Then, the generator must go through the codes to find the applicable provisions. The generator needs to interpret the provisions related to identifying the waste. As noted earlier, due to the complex nature of the regulations, the generator may not fully understand the regulation codes sufficient for waste identification, and usually relies on the service by TSDF for the waste identification task.

Determine the appropriate regulation code to comply with

Find the related provisions in the regulation code that can

be used for identifying a waste for a waste generator

Interpret correctly provisions that are needed for the waste

identification process

Apply provisions that are needed for the waste


Figure 4.3: The procedure for identifying a waste for a waste generator from a generator viewpoint

To the regulator, the procedure for identifying a waste using the regulation codes is more direct and explicit. The process is shown in Figure 4.4. The regulator, being the drafter of the codes, understands the structure of the codes, the terms used in the provisions, and the content of the regulation codes and the codes applicability. For example, when identifying a hazardous waste, a regulator knows that it is the 40 CFR 261, 262, and the related provisions in other parts of CFR (such as 40 CFR 268, 273 for possible exclusions or restrictions) that should be used. However, a regulator usually does not have the information on how a generator may use the regulation codes. While the complicated structure of a code may be easily understandable to a regulator, it may not be necessarily clear to a generator.

Use 40 CFR 260, 261, 262 for the hazardous waste compliance

standard applicable to generators

Use 40 CFR 262 subpart A section 262.11 as the beginning for

hazardous waste identification

Find the related provisions in 40 CFR 261 for details of

identification steps

Find the related exceptions of hazardous waste for a generator

Use provisions to determine if the waste is hazardous for a generator

Figure 4.4: The steps for determine a waste for a generator from a regulator viewpoint

Use 40 CFR 260, 261, 262 for the hazardous waste compliance

standard applicable to generators

Use 40 CFR 262 subpart A section 262.11 as the beginning for

hazardous waste identification

Find the related provisions in 40 CFR 261 for details of

identification steps

Find the related exceptions of hazardous waste for a generator

Use provisions to determine if the waste is hazardous for a generator

Determine the proper regulation code to comply with

Find the exact regulation code

Find the related provisions in the regulation code that can

be used for identifying a waste for a waste generator

Interpret correctly provisions that are needed for the waste


Apply provisions that are needed for the waste


TSDF’s

information and knowledge

for identifying a waste for a generator

Use 40 CFR 260, 261, and 262

Provide 40 CFR 260, 261, and 262

Provide the provisions to use, the

interpretations, and the procedure

for a waste identification

A generator’s knowledge

structure of identifying a waste

With the information and

knowledge from a TSDF,

a generator can obtain

sufficient knowledge for

using 40 CFR to identify

a waste for itself.

Figure 4.5: A TSDF provides a generator information and knowledge to perform waste identification

A TSDF is the participant that has both sufficient understanding of regulation codes and of when and how to use them for identifying the wastes. A TSDF knows what are the appropriate regulation codes and how to find them. A TSDF also knows which provisions to apply and how to interpret the related provisions. In addition, by working with the generators, a TSDF also has the knowledge about what the generators need to know for identifying wastes. From a functionality viewpoint, a TSDF is the mediator that tries to provide the information and knowledge to a generator so that a generator can perform the waste identification process as much as what a regulator would require a generator to comply. The interaction between a generator and a TSDF is shown in Figure 4.5.

In summary, the information and knowledge of a participant in the regulation compliance checking process has three aspects: (1) the characteristics of the information and knowledge source, including the structure or relations of the sources, (2) the content of the information and knowledge, and (3) the background knowledge for interpreting this particular information and knowledge content. They form a local context for representing the information and knowledge for a compliance checking for a certain participant. Figure 4.6 depicts the local context structure for each participant, a generator, a TSDF, and a regulator, respectively.

A Local Context

Aggregation relation:

A

contains

B

A

B

Knowledge Source

Background Knowledge

Knowledge Content

Figure 4.6: A context of information and knowledge structure for a participant

4.3.1.2 Characteristics of a Participant’s Compliance Knowledge

Using the context structure for organizing the information and knowledge for a compliance participant, the following properties can be stated.

1. the structure of local knowledge content: in the local context of a participant, the structure or relation of the content of knowledge is understandable with the domain specific background knowledge. For examples, a regulator knows certain provisions in certain regulation codes that are written for identifying a generator's waste. A generator knows the fact that its wastes must be identified. A TSDF knows how to provide a generator the proper procedure for identifying a waste using the proper regulation codes.

2. the background knowledge: the background knowledge provides the available resources within a context for interpreting the meaning of the knowledge source and the knowledge content. For example, a regulator has the background knowledge for interpreting all the regulation codes it drafts for hazardous waste identification for a generator, but a generator may not have sufficient background knowledge for using the regulation codes for identifying a waste. A TSDF has both the background knowledge of interpreting regulation codes and the background knowledge for guiding a generator as how to use the regulation codes for the waste identification.

3. the autonomy of the local knowledge content: within a local context, the content of the knowledge is assumed to be autonomous and self-explanatory since the background knowledge is given. The local context needs to be augmented with background knowledge to form a complete content of the knowledge. This point is important, especially in a distributed information and knowledge domain, where multiple local contexts have to go beyond its own context and interact with one another to obtain the background knowledge. For example, in the context of information and knowledge of a generator, it cannot understand the knowledge content of the regulation codes since it does not have the sufficient background knowledge as a regulator. But a generator knows the fact that it should contact its TSDF to obtain the additional background knowledge that is necessary for understanding the regulation codes.

4.3.2 Organization of Participants

A Context of Information

and Knowledge for a Regulator


A

contains

B

A

B

Knowledge Sources:

The repository for

regulation codes

Background Knowledge :

The implicit assumptions

and knowledge used in

writing codes and for

interpreting codes

Knowledge Contents:

The content of the

Regulation provisions

within the codes and

their correct

interpretations

Figure 4.7: The information and knowledge organization for a regulator

In our definition, a context for a regulator contains the following elements. (1) the knowledge source: the repository for the regulation codes, (2) the knowledge content: the content of the regulation codes and their interpretations, and (3) the background knowledge: the implicit assumptions and knowledge for the interpretation of regulation codes. Within the regulation compliance checking, a regulator also performs the functions for publishing regulation code, for revising regulation codes, and for enforcing compliance checking. The context of information and knowledge for a regulator is illustrated in Figure 4.7. As noted, the detailed discussion for organizing the information and knowledge in the regulation code is presented in Chapter 2.


and Knowledge for a Generator


A

contains

B

A

B

Knowledge Sources:

(1) The for available

regulation codes and

(2) the service from a

TSDF for interpreting

codes and compliance

procedure


Insufficient background

knowledge for interpreting

regulation codes for

waste compliance procedure

Knowledge Contents:

Limited knowledge

about regulation codes and

the procedure for using

codes for hazardous


Figure 4.8: The information and knowledge organization for a generator

For the compliance checking, a context for a waste generator contains the following elements. (1) the knowledge source: the available original, but not yet interpreted regulation codes and the service from a TSDF's information and knowledge base, (2) the knowledge content: the partial knowledge about regulation codes and the partial information and knowledge of its own wastes, and (3) the background knowledge: insufficient background knowledge for interpreting regulation codes and conducting waste compliance checking. In the regulation compliance checking, a waste generator has the following functions: the function for reporting waste information, the function for cooperating with its facilitators to perform compliance for the wastes, and the function for getting the compliance results. The context of information and knowledge for a generator is illustrated in Figure 4.8.


and Knowledge for a TSDF


A

contains

B

A

B

Knowledge Sources:

(1) The for available

regulation codes and

(2) Knowledge of how

to provide service to

generators


Sufficient background

knowledge for interpreting

regulation codes for


Knowledge Contents:

Sufficient knowledge

about regulation codes and

the procedure for using

codes for hazardous


Figure 4.9: The information and knowledge organization for a TSDF

The context for a facilitator includes (1) the knowledge source: the original regulation codes from regulators, cases and experiences from previous business practices, and available regulation interpretation and knowledge for conducting compliance checking, and (2) the knowledge content: the expertise for interpreting the regulation codes for compliance checking and applicability of certain regulation codes, and (3) the background knowledge: knowledge for correctly interpreting regulation codes, and the conditions for applying the codes. The context of information and knowledge for a regulator is shown in Figure 4.9. In regulation compliance checking, a TSDF plays the roles for interpreting regulation codes and their applicability for the generators, for checking compliance for the generators using the regulation codes from regulators, and for reporting the checking result back to the generators and to the regulators.

The context of the information and knowledge for the three participants is summarized in Table 4.1.

Components of Context

Knowledge Source

Knowledge Content

Background Knowledge

Participant

Regulator

The repository for regulation codes.

The content of the regulation codes and their correct interpretations.

The implicit assumptions and knowledge used for writing and interpretation.

Generator

The available original regulation codes, the service from TSDF for interpreting codes and compliance procedure.

The partial knowledge about regulation codes and the wastes produced.

Not enough background knowledge for interpreting regulation codes and waste information.

TSDF

The original regulation codes from regulators, regulation compliance checking cases from former business practices, and available regulation interpretation and knowledge for conducting compliance checking.

The expertise for interpreting the regulation codes for complicated checking process and applicability of a certain regulation code.

Sufficient knowledge for correctly interpreting regulation codes, and the conditions for applicability.

Table 4.1: Summary of components of context for compliance checking participants

To successfully execute compliance checking requires the involvement of multiple local contexts. The reason is that a local context usually does not contain sufficient information and knowledge for its own functions. In order to use the information and knowledge flow for regulation compliance, we need to include the contexts for coordination and cooperation in addition to the three local contexts. Figure 4.10 shows a detailed description of the local contexts for the three participants and their relations for information and knowledge interaction during a compliance checking process.

Context for TSDF

Context for Regulator

Context for Generator

Context

for Interaction

between Generator

and Regulator

Context

for Interaction

between Generator

and TSDF

Context

for Interaction

between TSDF

and Regulator

Information and

Knowledge Flow

Context

Figure 4.10: Information and knowledge interaction during compliance Process

As shown in Figure 4.10, it is the contexts about the interaction between the participants that connect the information and knowledge flow for each individual participant. To allow information and knowledge coordination in regulation compliance, the interaction and the interoperation for local information and knowledge context of participants are important.

4.3.3 Interoperation for Participants

To coordinate the information and knowledge from individual participants during compliance requires the determination of the relations among the local contexts of the participants and their interoperations. The relations include not only the individual contexts but also the interaction. The distributed information system for regulation compliance checking is thus a combination of the multiple local contexts for participants together with the meta level information regarding for interoperability among the local contexts.

4.3.4 Hierarchical Organization of Compliance Knowledge

The overall organization of a regulation compliance checking process can be structured as a hierarchy of knowledge base. The knowledge sources (KS) and knowledge contents (KC) in the local contexts are the physical entities that need to be interoperable with one another to fulfill the functions within the regulation compliance checking. The hierarchy of the distributed information and knowledge during compliance checking can be formally organized as a context graph structure shown in Figure 4.11.

Context for

Generator

Context for

TSDF

Context for

Regulator

KS

KC

KC

KC

KS

KS

Context for

Interoperability

between Generator

and TSDF

Context for

Interoperability

between Regulator

and TSDF

Context for

Interoperability

between Regulator

and Generator

Level 1

Context for Information and Knowledge Flow

during Compliance Checking Process

Level 2

Level 3

Figure 4.11: Hierarchy of contexts for information and knowledge flow

Level 1 of the hierarchy contains the local information and knowledge organization for individual participants; Level 2 contains the meta level information to enable interoperability among the local contexts; and finally Level 3 contains the knowledge for distributed information flow during regulation compliance checking.

Through the information and knowledge interoperation between a generator and its TSDF, the generator can obtain sufficient knowledge for using 40 CFR to identify a waste for itself. It is the contexts for individual participants (Level 1 context) together with the information and knowledge interoperability contexts between local contexts (Level 2 context) that provide sufficient information and knowledge for a generator to fulfill the task of regulation compliance for the waste identification process.

The hierarchy of contexts of information and knowledge provides the procedure for requesting and obtaining the necessary information and knowledge for a regulation compliance process, as shown in Figure 4.12. The procedure for context based information and knowledge interaction is based on two criteria: (1) using the information and knowledge from the nearest local context and (2) using the information and knowledge that is the most trustful one. This means that we don't need to make an assumption that the nearer the context is, the most trustful it is. This assumption has been used either explicitly or implicitly by other distributed information and knowledge organization approach, especially the ones that discuss the interoperability of information and knowledge [McCarthy, 1994; Wiederhold, 1994; Maluf and Wiederhold, 1997].

Procedure for InformationKnowledgeInteraction(input: compliance_process)

For the compliance_process, searching the related information and knowledge in the local context.

If (information found)

checking for the information and knowledge validness and completeness.

Else

traversing to next context level to find the information in the context.

If ( the information and knowledge is valid and complete)

quitting the InformationKnowledgeInteraction process.

Else

traversing to the next context level for finding the information in the context

Figure 4.12: Procedure for information and knowledge interaction based on contexts during compliance checking process4.3.5 Examples

To illustrate the application of using context based information organization and interoperation, we provide examples for resolving vagueness in interpreting regulation codes. For example, the definition of a hazardous waste is usually not clear in the context of information for a generator, or more specifically, the information and knowledge resources for the definition of a hazardous waste is basically incomplete. On the other hand, a TSDF usually knows the definition of a hazardous waste. Therefore, a TSDF can always provide a generator more accurate information about the compliance checking. However, the problem is that the way a generator handles the information is usually quite different from the way a facilitator handles the information, although they deal with the same problem. To be more specific, there are cases where (1) same term can have different meaning in different local context, and (2) different terms can have the same meaning in different local context.

For the case of the same term having different meaning, we use the following example to illustrate. The term “solid waste” may mean “the waste that is solid” to a generator, but the same term “solid waste” may include the waste that is not necessary in a solid form to a TSDF or a regulator. To make the discussion easy to read, we denote the symbol C("participant") to mean a context of information and knowledge of a participant in compliance process, the symbol C("participant"): Definition("key word") := "the definition content" to represent that in the context for a participant, the definition of the "key word" is "the definition content". For the case of the definition of "solid waste", we have:

C("generator"): Definition(“solid waste”) := “the waste that is solid”

C("TSDF"): Definition(“solid waste”) := “the waste that is defined in regulation codes, using either federal code 40 CFR 261 or a state specific code, or both”

C("Federal EPA"): Definition(“solid waste”) := “the waste that is defined in 40 CFR 261”

For the case of the different term having the same meaning, we provide another example below. For a certain regulation code, say the federal code for hazardous waste regulation applicable to waste generators, a generator may refer to it as the “Code of Federal Regulation chapter 40 part 262” in its information resource, but a TSDF may call it “40 CFR 262” in its information resource. Using the same notation as the first example, we have:

C("generator"): Definition(“Code of Federal Regulations chapter 40 part 262”) :=”the federal code applicable to generators”

C("TSDF"): Definition(“40 CFR 262”)

:=”the federal code applicable to generators”

C(“generator”)

Definition(“solid waste”)

= “the waste that is in a

solid form.”

C(“TSDF”)


= “the waste that is defined

as solid waste in 40 CFR

261 or in State regulation,

or both..”

C(“Federal EPA”)


= “the waste that is defined

as solid waste in 40 CFR

261.”

Resolution mechanism:

(1) form a unique term of

“solid waste”

(2) search for the most trustful definition of “solid waste” in all the local contexts

of

C(“generator”),

C(“TSDF”),

C(“Federal EPA”)

(3)

Lifting the most trustful one in

C(“TSDF”) as the definition of “solid waste”

The interoperable information and knowledge of C(“generator”), C(“TSDF”),

and C(“Federal EPA”)

Definition(“solid waste”) = “the waste that is defined as solid waste in 40 CFR 261 or in

State regulation, or both..”

Obviously, the information and knowledge mismatch cannot be resolved in any of the local context alone, since none of the local contexts has the complete information about the other. Only by organizing an interoperable meta information resolution mechanism for different local contexts can we resolve the local level information and knowledge mismatch. The following general principle is used to form interoperability among contexts for the above two examples: we first establish a global view of the information and knowledge in an interoperability context for individual contexts, and then match and combine the information in the interoperability context.

Figure 4.13: Using context lifting to resolve definition content conflicting from different contexts

For the first example, i.e., the case of using same semantic term that have several different contents, we provide the following mechanism, as shown in Figure 4.13, that resolves the conflicting of the definition of “solid waste" in different local contexts. We observe that the term “solid waste” is defined in a local context and its exact meaning varies from one local context, say C("Generator"), to another local context, say C("Federal EPA"). The knowledge difference is a reality in distributed local contexts and will lead to wrong solution during a regulation compliance process. The solution for this problem is to define a global view of the term “solid waste” that has the exact meaning in the regulation compliance process. The way to unify the meaning of the term “solid waste” depends on the decision of choosing the most reliable concept for “solid waste”. It turns out that in reality, usually the TSDF knows what is the exact meaning of the “solid waste” because of the experience they have. Therefore we use the definition of “solid waste” from the local context of the facilitator as the global concept of “solid waste” to all the other local contexts. We can use a denotation C("regulation compliance"): Definition( ) to define the concept “solid waste” in the context for regulation compliance checking, i.e., the context domain for regulation compliance, and thus provide the reliable knowledge for all the local contexts:

C("regulation compliance"): Definition(“solid waste”)

:= “the waste that is defined in regulation codes, using either federal code 40 CFR 261 or a state specific code, or both.”

After combining the definition into the global view, we arrive at a concept of “solid waste” that can be used in the regulation compliance domain as shown in the Figure 4.13.

For the second example, i.e., the case of using different terms to represent the same physical entity, we provide the following mechanism, as shown in Figure 4.14, that unifies the two different semantic terms with the same concept in two local contexts. What we need here is a mapping of the concept in local contexts into a unified global namespace. The unified namespace acts as a mediator in the form of a “multi-language dictionary” for local contexts, where the information in a local context use its own “language” for communication, but the terms used in different local contexts need to use the unified namespace to match the terms that have the same meaning. For this example, we can define the mapping below to show that the term “Code of Federal Regulations chapter 40 part 262” is the same as the term “40 CFR 262” in a regulation compliance checking process.

C("regulation compliance"):Definition(“Code of Federal Regulations chapter 40 part 262”)

:= ”the federal code applicable to generators”

C("regulation compliance"): Definition(“40 CFR 262”) )

:=”the federal code applicable to generators”

C(“generator”)

Definition(“Code of

Federal Regulations

chapter 40 part 262”)

= “the federal code

applicable to generators.”

C(“TSDF”)

Definition(“40 CFR 262”)

= “the federal code

applicable to generators.”

Resolution mechanism:

(1) search for the definition of “the federal code applicable to generator” in all the

local contexts of

C(“generator”),

C(“TSDF”)

(2) form an equivalence for the two definitions from two local contexts

(3)

Provide a Unified namespace that hosts the equivalence

The interoperable information and knowledge of C(“generator”), C(“TSDF”)

C(“generator”):Definition(“Code of Federal Regulations chapter 40 part 262”) =

C(“TSDF”):Definition(“40 CFR 262”) = “the federal code applicable to generators”

We believe that the above examples show the need for providing interoperable meta information, or interoperable context, for the information and knowledge in local contexts for regulation compliance process. On the other hand, they also reveal the fact that only in the interoperability context, not in any local context, can we organize the information and knowledge that can be used in the regulation compliance process.

Figure 4.15: Using unification to resolve definition key work conflicting from different contexts

Another implication of the context based information and knowledge organization is that the local context that contains the richest information and knowledge should be considered as the mediation base for other local contexts. The direct application of the idea is that we should consider organizing a mediation based computing paradigm around the context for facilitator, since the context for facilitator is the richest among the three contexts involved.

The organization of contexts and their interoperations to form the realistic information and knowledge flow for compliance checking offers a way for supporting modularity in system implementation. Next, we map the context based information and knowledge organization to a mediation based distributed computing paradigm [Wiederhold, 1994] to build a distributed information management system for hazardous waste regulation compliance process.

4.4 A Distributed Computing Paradigm

Mediation is a technology that is intended to provide a scalable information and knowledge integration in a multiple information resource environment [Wiederhold, 1994; Wiederhold and Genesereth, 1997]. The purpose of the mediation is to scale local domain based or context based knowledge so that the information and knowledge from many heterogeneous local sources can contribute to a certain application [Maluf and Wiederhold, 1997; Wiederhold and Jannink, 1999].

In this section, the topic of using mediation based information and knowledge integration for supporting context-based information and knowledge interaction for regulation compliance checking is discussed. The topic is divided into two subtopics. They are: (1) the topic of mapping from context based information and knowledge organization in regulation compliance process to a mediation based information and knowledge organization paradigm and (2) the topic of using distributed information and knowledge infrastructure to support mediation in regulation compliance process.

4.4.1 Mediation Based Information Organization for Compliance Process

We need distributed information integration model to facilitate the distributed information and knowledge interaction modeled by contexts. Here, we use the method of mediation as a coordination and cooperation paradigm to support the context based information and knowledge organization for the regulation compliance process.

Knowledge source

Knowledge content

Context for a participant

A directory for the

knowledge source

Organized knowledge

contents are indexed

by the directory so

that the knowledge

can be retrieved

root

Original

content

Explanation

Directory and Organized Information and knowledge for a participant

Original

content

To realistically form an integration model for context-based information structure, we need mappings for transforming the distributed contexts into tangible information and knowledge entities. Basically, we need mappings (1) from the local contexts of participants to a information and knowledge representation for the participants, and (2) from the interoperability contexts to the mediation based meta information structures.

Figure4.15: Directory and organized information and knowledge for a participant in regulation compliance process

First, a procedure, as shown in Figure 4.15, is employed to map the elements of a local context to an organized information and knowledge service in a local participant domain. The organized information resources can be the local information directories together with the enriched (that is, raw information that are interpreted) and well organized (that is, the one that has a schema) information content. The mapping from a local context of information to a structured local information resource for a participant can be conducted in the following steps.

1. Organizing the information sources to reveal the available resources.

2. Organizing the information content to make the implicit assumptions explicit and provide sufficient explanation for original information whenever necessary.

3.

Knowledge source

Knowledge content


A directory for the

knowledge source

Organized knowledge


by the directory so

that the knowledge

can be retrieved

root

Original

content

Explanation


Original

content

Knowledge source

Knowledge content


A directory for the

knowledge source

Organized knowledge


by the directory so

that the knowledge

can be retrieved

root

Original

content

Explanation


Original

content

Knowledge source

Knowledge content


A directory for the

knowledge source

Organized knowledge


by the directory so

that the knowledge

can be retrieved

root

Original

content

Explanation


Original

content

Meta

-information Directory of

the knowledge resources in

regulation compliance process

Mediation for information

interoperability

Setting up the hierarchical information identification directory for all the information resources and knowledge contents.

Figure 4.16: Meta-information directory and information interoperation for participants in regulation compliance process

The mapping should also include the tasks for classification of the related information and knowledge, the interpretation of the vagueness in information and knowledge, exception handling, the directory for querying other contexts when necessary. The mapping of interoperable context to meta information control, and information matching mechanism are given as shown in Figure 4.16. Meta information control consists of a directory for finding information about different participants and their meta-information in a regulation compliance process. Meta-information refers to the organization information about the information in local contexts. Meta-information is related to the hierarchical organization of contexts of information and knowledge organization. To fulfill the task of building interoperability contexts among the local context of information and knowledge, we can follow the steps below.

· Identifying the necessary information and knowledge for communication among local contexts.

· Building up the interoperability directory for communications among the local contexts.

· Building the matching rules, whenever necessary for resolving information mismatches.

The need for rules for resolving information mismatching is to deal with the information and knowledge interoperability among the local contexts.

In the above mappings from the context based information organization to the mediation based information infrastructure, we view a local context as a piece of distributed knowledge, while we view an interoperability context as a globally available interoperable knowledge supported by a trustful mediator. In this sense, a mediator of regulation compliance serves as the global information and knowledge coordination center for a distributed information and knowledge management for regulation compliance process.

4.4.2 Supporting Distributed Information and Knowledge Interoperability

The information organization in regulation compliance process can be supported by a mediation based distributed information model. Next, we develop the necessary functional components to provide the formulation for supporting implementation for the mediation based model. We will employ the method that is developed in mediation [Wiederhold, 1994; Maluf and Wiederhold, 1997] to deal with forming functional components in regulation compliance process.

Forming a mediation based information management infrastructure for regulation compliance requires the functional components for

1. global view of the information and knowledge for information and knowledge interoperation in regulation compliance process,

2. information and knowledge management modules for local contexts, and

3. interfaces for information and knowledge matching and interchange for the local contexts.

First, we define the components for basic global views of the information and knowledge for composing compliance checking functions:

· Unified Definition Resource (UDR): the resource that unifies the local definitions in local contexts and provides a trusted resource for resolving any uncertainty about the definition. The definitions in UDR can be formed from the definitions in all the local contexts and it should also provide the meta-level rules for resolving conflicts in definitions. The meta-level rules usually provide the resolution through the principle of more specific and trustful definition overrides less specific and trustful definition. For examples, the detailed definition from a state regulation code overrides the definition in Code of Federal Regulations. The definition that is created more recently overrides the ones that are created less recently; for example, the definition used in regulation code revised in 1998 overrides the definition used in 1990 regulation code, and the definition that provides higher authority overrides the definition from the lower authority, for example, the exact definition provides by the regulator overrides the same definition in a certain generator’s interpretation document. The definition conflicting examples depicted in Figures 4.13 and 4.14 show the need for this component.

· Unified Information Resource Locator (UIRL): there are knowledge resources, knowledge contents, and transferable background knowledge in local contexts. The information and knowledge is organized in a hierarchical structure. Therefore, we need a information resource locator to find out the location for a specific information and knowledge in the hierarchy. This locator also keeps the meta-information about the overall information organization of the regulation compliance process.

Second, we provide the generic functional components for composing information and knowledge from local contexts

· Information and Knowledge Lifting (IKL): IKL is the function for exporting the information and knowledge in a local context to the global information resource. This function is useful because we need reliable and sound information and knowledge resources that can be used by the compliance process. For example, the information and knowledge of the original regulation codes from the regulator can be exported to the whole regulation compliance domain as the regulation code repository for compliance checking.

· Information and Knowledge Intersection (IKI): IKI is the function for combining the minimum amount of information and knowledge that are necessary for a functional context. For example, during the process for regulation code revision, we can use the concept of IKI to express only the changed provisions for the revised codes and the old codes. As a result, the regulator and the facilitator can focus on revising only the information and knowledge representation for the changed provisions and thus reduce the reengineering task for revising regulation codes.

· Information and Knowledge Union (IKU): IKU is the function for combining the information and knowledge in the related local contexts. One main application of this function is the process of forming the unified definition resource in the regulation compliance domain. The operation of this function creates a consistent new global definition resource that can be used by all the functional components in compliance checking.

· Information and Knowledge Difference (IKD): IKD is the function for removing the shared identical information and knowledge from the related local context. This operation can find its application in describing the imbalance of the background knowledge in different local contexts. For example, after we find the knowledge differences between a generator and its TSDF, we can provide information and knowledge exchange that is needed for interpreting the provisions in regulation codes.

· Information and Knowledge Exception Handling (IKEH): IKEH is the function for dealing with unusual information and knowledge. For example, during a regulation compliance checking, there are cases that do not have enough information available in the input data, and therefore the checking process must be able to deal with the incomplete information. The need for exception handling is the requirement for dealing with non-monotonicity [McCarthy, 1980] in the information and knowledge organization in compliance checking process.

By implementing these functional components, we have the tools for building a mediation based information and knowledge management system for hazardous waste regulation compliance checking. Details about the implementation infrastructure for mediation-based system for supporting compliance checking will be presented in Chapter 6, where the implementation of a prototype system is described.

4.5 Related Research

The Context based information and knowledge organization approach is related to various methods used in the distributed information and knowledge paradigm. Among them are the research in formalizing context [McCarthy, 1993], mediation and its related ontology [Wiederhold, 1994; Wiederhold and Genesreth, 1995; Maluf nad Wiederhold, 1997], knowledge and meta knowledge in distributed systems [Halpern and Moses, 1990; Barwise and Seligman, 1997].

The formal theory of context argues that in a complex information and knowledge organization environment, the participants organize and use knowledge for their own purpose in their individual subcontext. As a result, the information and knowledge created from the subcontext is not ready for direct sharing in a more general context [McCarthy, 1994]. Therefore, we need to make explicit the individual context when sharing information among multiple participants. This is a common situation that the theory of context tries to deal with. In the hazardous waste compliance process, the information and knowledge for each participant has its own local context and in order to make the local contexts work together, we must explicitly formalize the local contexts together with the interoperability contexts. We employ the basic concept of context from the formal theory of context, and furthermore, we give an explicit formulation of the distributed contexts in the compliance checking process.

Mediation and the related algebra over ontology provide intermediary services in distributed information systems [Wiederhold, 1994; Wiederhold and Genesereth, 1995; Wiederhold and Jannink, 1999]. The need for mediation is the result of dealing with only the interrelated information, not the complete information in a distributed system. Mediation and the related algebra over ontology also provide a formal methodology for organizing the interoperability of information. There are many applications that employ mediation as a major tool for distributed information system design [Wiederhold and Genesereth, 1995]. But the application of mediation in regulation compliance checking process is a new attempt. We use mediation for formulating a distributed computing paradigm for regulation compliance checking process, particularly, we use it for organizing distributed information and knowledge infrastructure in compliance checking. This infrastructure can be used to provide guidelines for the implementation of a distributed information management system for compliance checking system.

Knowledge and meta knowledge is a research topic in the theory of distributed systems [Halpern and Moses, 1990; Barwise and Seligman, 1997]. In our research, the topic motivates us to investigate what is the explicit information and knowledge globally known in a distributed compliance checking process and what is the hidden knowledge that is used to interpret the explicit information and knowledge. The distinction between the knowledge and hidden knowledge is of particular importance in forming a context of information and knowledge for a participant during a compliance checking. The explicit treatment of the background knowledge in a context is largely originated from the treatment of common knowledge in a distributed information system. In addition, the formulation of the distributed information interoperation in a distributed compliance checking process is also an attempt for reaching an agreement for using the common knowledge for each participant involved in compliance process.

4.6 Summary

In this chapter, we provide a methodology for organizing distributed information and knowledge. The critical step is the forming of contexts for information organization for the participants in hazardous waste regulation compliance process. We begin with a detailed analysis of the process for conducting hazardous waste regulation compliance and the related distributed information and knowledge during the process. Then, we define the concept of context of information organization and apply the concept to the local contexts for waste generators, facilitators, and regulators. Then, we define the meta information organization for making the local context interoperable. We introduce the meta information for organizing the hierarchy for searching information and knowledge through local contexts using a combination of criteria of the most trustful and the nearest information and knowledge sources and contents. Finally, we map the context based information organization for regulation compliance process to a distributed computing paradigm using a mediation based method. In this way, we use the methods developed in the mediation methodology to study the information and knowledge representation and information interoperability in the practical compliance checking process.

In summary, we use the context based information organization to provide the foundation for the distributed information and knowledge management for the hazardous waste regulation compliance process. In addition, we use a mediation based distributed computing paradigm to provide the implementation guidelines.

chapter 4 distributive regulation frameworkeil.stanford.edu/publications/shawn_kerrigan/temp... ·...

Documents