Multispace search for minimizing the maximum nodal degree

Hajek and Sasaki (1988) showed that, for continuous traffic and packet radio network, the selection of paths that minimize the maximum nodal degree generates schedules of minimum-length. This result suggests that minimization of the maximum nodal degree provides good (although not necessarily optimal) performance in slotted networks with fixed-length packets. We give a multispace search algorithm that interplays structural operations in conjunction with a local search algorithm for the minimization of the maximum nodal degree. Structural operations disturb the environment of forming local minima, which makes multispace search a very natural approach to the problem. Experimental results indicate that this method has improved local search in terms of the solution quality and its sensitivity to the initial random assignment.