A knowledge-based expert system (kbes) approach can be used to solve the single-mode (automobile), fixed-demand, discrete, multicriteria, equilibrium transportation network design problem. previous work on this problem revealed that mathematical programming methods perform well on small networks with only one objective. a solution technique is needed that can be used on large networks that have multiple, conflicting criteria with different weights of relative importance. the kbes approach discussed in this paper represents a new way to solve network design problems. the development of an expert system involves three major tasks: knowledge acquisition, knowledge representation, and testing. for knowledge acquisition, a computer-aided network design and evaluation model (ufos) was developed to explore the design space. this study investigated the problem of designing an optimal transportation network by adding and deleting capacity increments to or from any link in the network. three weighted criteria were adopted for use in evaluating each design alternative: cost, average volume-to-capacity ratio, and average travel time. the best nondominated design is determined by a multicriteria evaluation technique called concordance analysis. the research started with a design exercise conducted by a group of students who were asked to find a series of link capacity changes that would produce a series of successively better designs. the results were carefully examined and used to generate the facts and rules that make up the knowledge base of the network design expert system (expert-ufos). it has two phases of analysis. the macrolevel analysis recommends a total budget using trade-off functions for each pair of criteria. the microlevel analysisprovides advice about how to add or delete capacity on each link toavoid paradoxes. test results show that expert-ufos found, with fewer design cycles, designs that were better than any of the 76 student designs included in the test. expert-ufos may have enough simplicity to deal with large networks. the results of this study, in which a laboratory-based knowledge acquisition method was employed successfully to generate a functional knowledge base, suggest that the kbesapproach is an appropriate method for dealing with the computational complexities of network design problems. this paper appeared in transportation research record no. 1145, expert systems for transportation applications. f:or covering abstract see irrd no 817754.
Abstract