Solutions to Resource Constrained Multi-Project Scheduling Problem (RCMPSP) were traditionally based on priority rules (PR) heuristics, and meta-heuristic approaches such as Genetic Algorithms (GA) and Evolutionary Strategies are recently proposed. In order to improve performance of these solutions, this paper solves the centralized RCMPSP using Priority Rules (PR) heuristics as best-case scenario for GA initial population. Using recently improved heuristics, this research analyze 12 well known priority rules on 14 different sets of simultaneously scheduled multi-projects commonly found in literature, with some generalized characteristics such as activity on node, deterministic durations, limited global resources and minimum Total Makespan (TMS) as objective. Furthermore, in order to evaluate the initial population performance, experimental results are compared, in terms of TMS and average project delay (APD). Analysis results shows that GA approach can lead to better solutions in some cases, in terms of TMS and APD, using an improved initial population.