<正> This paper describes an application of using Genetic Algorithms (GA) as an optimization tool to determine the optimal multiyear work plan for a pavement network. Nearly every highway agency must face the problem of developing a multiyear work plan for the pavement network under its jurisdiction. Pavement conditions within the network vary with time and the treatments received previously. As the size of pavement network increases and the number of treatment options grow, intuitive, heuristic, or "worst-first" approach to determine the work plan may not result in the most effective way to spend the available budget. Modern pavement management information systems (PMIS) typically contain large amount of data (e.g., section location, length, width, conditions, etc.) that can be used for such decision-making. The problem of finding the best multiyear work plan can be modeled as a combinatorial optimization problem. The objective may be to achieve the highest possible level of average network conditions for the given budgets. Other constraints that could be incorporated include geographic, institutional, and political. The solutions obtained from GA method were verified with that from dynamic programming method, a traditional optimization technique. The major benefits of using GA for solving the combinatorial optimization problems are its flexibility and scalability. GA technique is flexible in that the constraints can be easily added, removed, or modified. Once the problem has been formulated, the size of the problem can be easily changed, although larger problems would take significantly more time to solve, as the solution space increases exponentially. In contrast, dynamic programming solutions are typically problem-specific. Therefore, they are very difficult to modify and become prohibitively complex for large problems. The GA based optimization technique was implemented using VBA in a spreadsheet and applied to the pavement network of the City of Toledo, Ohio. The GA optimization module is a part of the pavement management information system developed for the city. A knowledge-based system was used to limit the size of the solution space prior to optimization. It reduces the number of possible multiyear treatment combinations for each road section. The result is a significant reduction in time required to find the GA solution.