Multiobjective hierarchical 2G/3G mobility management optimization: niched Pareto genetic algorithm

We first propose four-layer optimization for a UMTS coverage area: (i) cell-oriented intra-SGSN (serving GPRS service node) layer, which is optimized RAs (routing areas) covering the intra-SGSN signaling cost, paging cost and RA load balancing; (ii) RA-oriented intra-MSC (mobile switching centre) layer, which is optimized location areas (LA) covering the intra-MSC signaling cost and LA load balancing; (iii) RA-oriented inter-SGSN layer, which is optimized SGSN coverage areas covering the inter-SGSN signaling cost, RNC (radio network controller) and SGSN load balancing; (iv) LA-oriented inter-MSC layer, which is optimized MSC coverage areas covering the inter-MSC signaling cost and MSC load balancing. We focus on RA optimization, namely layers (i) and (iii). The optimization of MSC coverage areas and LAs is performed in a similar manner. We propose a schema-based niched Pareto genetic algorithm, which deals with multiple objectives by incorporating the concept of Pareto domination in its selection operator, and applying a niching pressure to spread its population out along the Pareto optimal tradeoff surface. The proposed genetic algorithm uses a schema-based partially matching crossover using tournaments of n size, where the crossover pairs are chosen in two steps, first based on the class ranking and then schema ranking. New offspring are modified using the geographical footprints to converge to the optimal solution faster.