Constraint transformation algorithm

The constraint transformation algorithm accepts a network of goals partially ordered by constraints, and generates a constraint network of primitive actions, such that, if there exists a directed path from goal A to goal B (i.e., A must be achieved before B) in the first network, and if OP-A is the primitive action that achieves goal A, and OP-B the action that accomplishes B, then OP-A and OP-B are labels on nodes in the generated network, and there exists a directed path from the node labeled with OP-A to the node labeled with OP-B. 

This separation of the variables allows a vast reduction in the number of calculations. This savings becomes more significant as the number of data points increases. Further, note that the final summation in this equation is in the form of a one-dimensional Fourier transform. This implies that the considerable calculational advantage of the fast Fourier transform (FFT) algorithm may be used here. The entire summation may be performed by repeated application of the one-dimensional FFT. This implies that for any data set that it is practical to apply the FFT, it would be also practical to apply the nonlinear constraints for improvement. 

The basic idea in OA/ER is to relax the nonlinear equality constraints into inequalities and subsequently apply the OA algorithm. The relaxation of the nonlinear equalities is based upon the sign of the Lagrange multipliers associated with them when the primal (problem (6.21) with fixed y) is solved. If a multiplier A is positive then the corresponding nonlinear equality hi(x) = 0 is relaxed as hi x) <0. If a multiplier A, is negative, then the nonlinear equality is relaxed as -h (jc) < 0. If, however, A = 0, then the associated nonlinear equality constraint is written as 0 ht(x) = 0, which implies that we can eliminate from consideration this constraint. Having transformed the nonlinear equalities into inequalities, in the sequel we formulate the master problem based on the principles of the OA approach discussed in section 6.4. 

Therefore, there is a penalty associated with the aforementioned transformations in the OA, OA/ER, and OA/ER/AP algorithms so as to treat general nonlinearities. It is, however, the GOA approach that represents an alternative general way of treating only inequality constraints. 

The resulting model is a MINLP with linear constraints and nonlinear objective function. The objective function terms can beconvexified using the exponential transformation with the exception of the condenser-cold utility expressions. If we replace the LMTDs with 2-3 times the ATmln, then, the whole objective can be convexified Floudas and Paules (1988). This implies that its global solution can be attained with the OA/ER or the v2-GBD algorithms. 

In contrast to the sequential solution method, the simultaneous strategy solves the dynamic process model and the optimization problem at one step. This avoids solving the model equations at each iteration in the optimization algorithm as in the sequential approach. In this approach, the dynamic process model constraints in the optimal control problem are transformed to a set of algebraic equations which is treated as equality constraints in NLP problem [20], To apply the simultaneous strategy, both state and control variable profiles are discretized by approximating functions and treated as the decision variables in optimization algorithms. 

Gautam and Seider (25) implemented the quadratic programming algorithm of Wolfe (30, 31) because compositions satisfy inequality constraints (n. 0) without transformations. Moles of solid... 

---