The optimization of operational states of various distributed power sources and energy storage devices is crucial for achieving efficient, economic, and reliable operation in photovoltaic-battery-direct current-flexible (PBCF) microgrids. However, this optimization process faces significant challenges due to dual uncertainties arising from equipment parameter configuration and operational power dispatch. This study investigates a PBCF microgrid system featuring a multi-layer DC bus architecture, integrating photovoltaic generation, energy storage systems, electric vehicles, central air conditioning units, and electrical loads. A multi-objective optimization function is formulated, incorporating operational cost, carbon emissions, and flexible load regulation as key performance indicators. To address these challenges, a multi-task evolutionary chaotic particle swarm optimization algorithm (MTECPSO) is proposed to enable simultaneous optimization of both equipment parameter configuration and power dispatch strategies. Simulation results from a case study indicate that, compared to conventional single-objective and multi-objective evolutionary optimization approaches, the proposed method achieves substantial improvements in computational efficiency while effectively satisfying the multi-objective optimization requirements.