Movable Antennas (MAs) and Reconfigurable Intelligent Surfaces (RISs) are two emerging technologies that enable dynamic reconfiguration of the wireless propagation environment. This paper investigates their joint use in a downlink multi-user MIMO system, where both the antenna positions at the base station and the RIS phase shifts are optimized to maximize the achievable sum-rate. The system model accounts for practical hardware impairments through an error-vector-magnitude (EVM) distortion formulation and employs linear zero-forcing precoding at the transmitter. The resulting design problem is highly non-convex with coupled continuous variables and stringent geometric constraints on the MA locations. To tackle this challenge, we propose an Enhanced Differential Evolution (EDE) algorithm that combines self-adaptive control parameters with a constraint-repair mechanism and a lightweight local search step, enabling efficient exploration of the feasible configuration space. Simulation results demonstrate a strong synergy between MA repositioning and RIS wavefront shaping: the proposed joint optimization framework significantly outperforms fixed-antenna and RIS-only benchmarks as well as a conventional PSO-based scheme, achieving improvement in sum-rate under representative system settings, while maintaining a moderate computational complexity suitable for offline or slow-timescale network reconfiguration.