This paper addresses the dimensional-synthesis-based kineto-elastostatic performance optimization of the delta parallel mechanism. For the manipulator studied here, the main consideration for the optimization criteria is to find the maximum regular workspace where the robot delta must posses high stiffness and dexterity. The dexterity is a kinetostatic quality measure that is related to joint's stiffness and control accuracy. In this study, we use the Castigliano's energetic theorem for modeling the elastostatic behavior of the delta parallel robot, which can be evaluated by the mechanism's response to external applied wrench under static equilibrium. In the proposed formulation of the design problem, global structure's stiffness and global dexterity are considered together for the simultaneous optimization. Therefore, we formulate the design problem as a multi-objective optimization one and, we use evolutionary genetic algorithms to find all possible trade-offs among multiple cost functions that conflict with each other. The proposed design procedure is developed through the implementation of the delta robot and, numerical results show the effectiveness of the proposed design method to enhancing kineto-elastostatic performance of the studied manipulator's structure.