Thermodynamic methods of process synthesis are very useful for the design of complex and energy intensive processes, but they cannot be used simultaneously with material balances. Algorithmic methods are simultaneous, but they are difficult to solve for complex and energy intensive processes because the number of variables increases with the number of combinations. We can approach the optimal design for complex and energy intensive processes if we combine the two methods. The combined approach is composed of two steps, the thermodynamic and the algorithmic one. In the first one we eliminate unpromising structures and we include new, potentially good ones by studying an Extended Grand Composite Curve. In the second one we can optimize the superstructure obtained by using Mixed-Integer Nonlinear Programming. The combined approach can be used for optimal design of energy and material parameters of continuous processes as well as for energy recovery. In a retrofit case study we have targeted energy saving using rigorous models and fixed amount how rates to find two promising structures, and then we have used parameter and simultaneous structural optimization to determine the best alternative and its parameters.