A genetic algorithm for privacy preserving combinatorial optimization

We propose a protocol for a local search and a genetic algorithm for the distributed traveling salesman problem (TSP). In the distributed TSP, information regarding the cost function such as traveling costs between cities and cities to be visited are separately possessed by distributed parties and both are kept private each other. We propose a protocol that securely solves the distributed TSP by means of a combination of genetic algorithms and a cryptographic technique, called the secure multiparty computation. The computation time required for the privacy preserving optimization is practical at some level even when the city-size is more than a thousand. Categories and Subject Descriptors H.4 [Information Systems Applications]: Miscellaneous General Terms Algorithms Keywords Privacy, TSP, GA, Privacy-Preserving Optimization

• Type: Conference
• Year: 2007
• Where: GECCO
• Authors: Jun Sakuma, Shigenobu Kobayashi
• Keywords: GECCO 2007, Genetic Algorithms, Secure Multiparty Computation, Traveling Salesman Problem