EXPERT SYSTEMS IN MUNICIPAL SOLID WASTE
MANAGEMENT PLANNING
By Brad Thomas
1
, Dave Tamblyn
2
, and Brian Baetz
3
INTRODUCTION
We readily enjoy the products of an industrialized society but are reluctant to accept that these increasingly
complex products are creating wastes at a continually increasing rate. A waste management plan is required to
optimize the allocation of public funds in a manner that best suits the local community and incorporates all
social and political considerations. Waste generation rates are increasing, existing landfills are approaching their
capacity, and the number of new recycling and treatment alternatives is increasing. In this setting, it is difficult
to evaluate existing facilities and establish new facilities that incorporate the best combination of waste
recycling, treatment, and disposal practices. There is clearly a need for solid waste management planning tools
that can be used to investigate the range of available alternatives and technologies and their associated
constraints.
In this note, we investigate the applicability of expert systems to waste management planning. An overview
of expert systems is provided, and the characteristics that suit their use in waste management planning are high-
lighted. Issues in solid waste planning are discussed, leading to a critical evaluation of the potential role of
expert systems in addressing these issues.
OVERVIEW OF EXPERT SYSTEMS
The advent of artificial intelligence technology, such as expert systems, may assist in
information/technology transfer in municipal solid waste management planning. To determine if this is practical,
we examined the characteristics of expert systems and the types of problems to which they are applicable.
An expert system is a "knowledge based system whose performance is intended to rival that of human
experts" (Lundberg and Robinson 1988). Expert systems consist of a knowledge base, an inference engine, and
an input/output mechanism (Myers 1986). A knowledge base is a set of true statements about a specific subject
or "domain." An expert systems engineer, or "knowledge engineer," gathers these facts or solution techniques
from a domain expert through communication about the problem to be solved. The user who wishes to solve a
problem enters data relating to the specific situation, and the expert system outputs the recommended course of
action to reach a solution or "goal state."
How does the expert system come up with its answer? It uses an "inference engine" that derives solutions
with varying levels of uncertainty from the data entered by the user and the data contained in the knowledge
base. Heuristics (logical deductions, rules of thumb, etc.) are stored in the knowledge base to guide the search
from a given state along the most probable route to the desired state. For example, a medical diagnosis expert
system would search through a knowledge base to arrive at a short list of illnesses that best fit the known
symptoms. In the field of waste management, the user might enter site-specific waste characteristics, and further
data such as costs and local regulations as required. After a search of possible waste reduction, treatment, and
disposal options, and perhaps interaction with existing "number-crunching" subroutines, the expert system
would identify possible courses of action; for example, one course may be to recycle 30% of the waste stream,
incinerate the rest, and dispose of the incinerator residues in an existing landfill.
1
Research Asst., Dept. of Civ. Engrg., McMaster Univ., Hamilton, Ontario, Canada L8S 4L7
2
Res. Asst., Dept. of Civ. Engrg., McMaster Univ., Hamilton, Ontario, Canada
3
Asst. Prof., Dept. of Civ. Engrg., McMaster Univ., Hamilton, Ontario, Canada.
Note. Discussion open until May 1, 1991. To extend the closing date one month, a written request must be filed with
the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on March
21, 1990. This paper is part of the Journal of Urban Planning and Development, Vol. 116, No. 3, December, 1990. p.150-
155. ©ASCE, ISSN 0733-9488/90/0003-0150. Paper No. 25370.