Flexibility of the energy system plays a vital role when integrating non-dispatchable renewable energy technologies. However, flexibility of the energy system has been often discussed only focusing on the operation of the energy system. This study extends the flexibility concept considering both design and operation of the energy system. In order to achieve this, pseudo chronological scenarios used for stochastic optimization is used to define system flexibility. Multiple criterions are considered when evaluating the flexibility of the system and fuzzy logic is used to consider the ambiguity in the assessment process when localizing into a specific application. Subsequently, multi objective optimization is conducted to design a multi-energy hub considering net present value (NPV), system flexibility and renewable energy generation. GPU-accelerated computing is introduced to speed up the computing when evaluating the objective functions for number of scenarios. Results of the study show that poor system flexibility can leads to poor utilization of renewable energy generated. More importantly, penetration levels of non-dispatchable renewable energy technologies notably reduce by 20-30% when considering the flexibility of the energy system which guarantees robust operation.