This thesis deals with the general nonlinear constrained optimization techniques. The developed algorithms can overcome the shortcomings of the reported techniques in classical and evolutionary optimization fields. A hybrid evolutionary programming (EP), a two-phase EP (TPEP) and an Evolian techniques are proposed. An improvement over the Evolian method is also presented. The first method, the hybrid EP, combines the evolutionary optimization approach with the classical gradient-based neural network method. The second method, the TPEP approach, is an extension to the hybrid EP, which does not require the gradient information. It utilizes an augmented Lagrangian function, and the Lagrange multipliers are updated in accordance with the constraint violation. The third method, Evolian, incorporates the concept of multi-phase optimization process and constraint scaling techniques to resolve the ill-conditioning problem. In each phase, the typical EP is performed using the augmented Lagrangian objective function, and the Lagrange multipliers, the constraint scaling factors, and the penalty parameters are updated at each phase transition. A novel method, Evolian II, is also proposed to improve Evolian with respect to solution feasibility, global convergence, computational complexity, and convergence speed. The interior and augmented Lagrangian penalty methods are combined to obtain a feasible solution trajectory with a moderate amount of computation time. A novel termination criterion is introduced in Evolian II, which reduces the computational burden in determining the convergence. It is shown that the optimality of the converged solution can be guaranteed with moderate assumptions. The use of subpopulation scheme and multiple Lagrange multipliers helps to improve the global convergence property of Evolian II.