Darrieus type hydrokinetic turbines (HKTs) with fixed pitch blades, like their wind turbine equivalents, exhibit poor starting torque whether blades are straight, troposkein or helical. High solidity improves starting torque to some extent but reduces peak efficiency. Straight blade turbines have been observed to shake violently due to cyclical hydrodynamic forces on blades. To solve this problem, the computer program VAWTEX-B was used to model blade forces and to predict how variable pitch might reduce shaking while maintaining strong starting torque and high peak efficiency C_pmax. A memetic algorithm was developed to search for optimum pitching parameters to achieve these objectives. Five optimization objectives were initially used, to maximize starting torque and C_pmax, to minimize downstream and the side force fluctuations causing shaking, and to minimize a function combining these four variables. A second optimization used a different objective function incorporating C_pmax and then minimizing a shaking force parameter while allowing a small reduction in C_Pmax. Modeling predicted reduced shaking while maintaining starting torque and peak efficiency well above those achievable with fixed pitch, and only slightly lower than the maxima achievable with variable pitch.