An online genetic algorithm for dynamic Steiner tree problem

Dynamic Steiner tree problem (DST) is described as the multipoint routing problem in the communication networks, in which a set of nodes and the set of Steiner vertices in the connection is changing with time. The problem is based on the Steiner tree problem on graphs. In the DST problem, the Steiner minimal tree cost is changing with time because the requests of the Steiner vertices or weights of graph edge are changing with time. We propose an online genetic algorithm (OLGA) to solve this problem. OLGA is different from general GA in that it does not evaluate individual against a fixed object, since the object changes with time. An individual in GA is expressed by the gene of the graph nodes. As an adapt method, the Prim algorithm is used to calculate the value of the cost of the graph nodes. In order to make our algorithm robust, we applied the B-problem set from Beasley (1990), thus the problems are based on the B-problem set. From experimental simulations, we show that the OLGA is robust to calculate this kind of problems