00822754

Component

https://doi.org/10.3141/1774-03

http://scholar.google.com...adek&publication_year=2001

A hybrid artificial intelligence approach based on combining simulated annealing (SA) and case-based reasoning (CBR) is presented. The approach is designed to allow for solving complex, time-critical optimization problems, examples of which lie at the heart of several intelligent transportation systems applications. According to this hybrid approach, the system, when faced with a new problem, first accesses the case base and attempts to locate a sufficiently similar case. If such a case can be located, the solution of the CBR-retrieved case is directly reused. If not, the SA algorithm is used to solve the problem, with the solution of the CBR-retrieved case serving as a starting point for the search algorithm. The rationale behind the approach is discussed, and its major design issues are analyzed. A case study is then presented to illustrate how the approach could be designed and to demonstrate its advantages. Results indicate that combining SA and CBR offers an efficient approach to solving complex, time-critical optimization problems. The results also indicate that SA, used in conjunction with CBR, should be started from a very low temperature to take advantage of the fact that the CBR-retrieved solution is close to the optimum.

This paper appears in Transportation Research Record No. 1774, Artificial Intelligence and Intelligent Transportation Systems.

English

p. 18-24

2001

Transportation Research Record

0309072352

Figures (4) ; References (12) ; Tables (2)

Artificial intelligence; Case based reasoning; Case studies; Optimization; Problem solving; Simulated annealing; Transportation

Planning and Forecasting; Transportation (General); I72: Traffic and Transport Planning

TRIS, TRB, ATRI

Jan 29 2002 12:00AM

• ADVANCED TECHNOLOGY AND INTEGRATED PUBLIC TRANSIT: SAN GABRIEL VALLEY SMART SHUTTLE FIELD OPERATIONAL TEST
• AUTOMATED VEHICLE IDENTIFICATION TAG-MATCHING ALGORITHMS FOR ESTIMATING VEHICLE TRAVEL TIMES: COMPARATIVE ASSESSMENT
• AUTONOMOUS AGENTS FOR TRAFFIC SIMULATION AND CONTROL
• BENEFITS EVALUATION OF BASIC INFORMATION DISSEMINATION SERVICES
• DEVELOPMENT OF CANADIAN ARCHITECTURE FOR INTELLIGENT TRANSPORTATION SYSTEMS
• EVALUATION FRAMEWORK FOR DYNAMIC VEHICLE ROUTING STRATEGIES UNDER REAL-TIME INFORMATION
• EVALUATION OF THE ENVIRONMENTAL EFFECTS OF INTELLIGENT CRUISE CONTROL VEHICLES
• IMPLEMENTING ADAPTIVE DRIVING SYSTEMS FOR INTELLIGENT VEHICLES BY USING NEURO-FUZZY NETWORKS
• MEASURING AGGREGATE PRODUCTIVITY BENEFITS FROM INTELLIGENT TRANSPORTATION SYSTEM APPLICATIONS: THE CALIFORNIA EXPERIENCE
• PEDFLOW: DEVELOPMENT OF AN AUTONOMOUS AGENT MODEL OF PEDESTRIAN FLOW
• TIME MANAGEMENT IMPACTS OF PRETRIP ADVANCED TRAVELER INFORMATION SYSTEMS: FINDINGS FROM A WASHINGTON, D.C., CASE STUDY
• USE OF FUZZY INFERENCE FOR MODELING PREDICTION OF TRANSIT RIDERSHIP AT INDIVIDUAL STOPS
• USE OF THE NATIONAL ARCHITECTURE TO DEVELOP AN INTELLIGENT TRANSPORTATION SYSTEMS STRATEGIC PLAN: CASE STUDY FOR A MEDIUM-SIZED AREA