Simplex defining

Any single point in the factor space may be defined in terms of the constituents of the mixture, Xj, Xj, Xj. All points within the L-simplex may also be defined, in exactly the same way, in terms of the three mixtures X/, X, X, at the vertices of the simplex defining the experimental domain. These mixtures X,, X2, Xj are termed pseudocomponents. The pure substances 1, 2, and 3 are the components. The conversion of system defined in terms of the pure components into one defined in these pseudocomponents is known as the L-transformation. The... 

Table 3.23 contains the experimental conditions for the initial simplex defined for our example and Figure 3.10 displays it in a two-dimensional space. 

Specifically, if the nuclear charge vector z falls on or inside of the Z-space simplex defined by the nuclear charge... 

It can be shown that this can be generalized to the case of more than two variables. The standard solution of a linear programming problem is then to define the comer points of the convex set and to select the one that yields the best value for the objective function. This is called the Simplex method. 

Kumiawan noticed that the first vertex was the same in both optimizations. This was due to the fact that in both cases the worse vertex was the same. Kumiawan also noticed that the search for the optimal conditions was more effective when two responses were optimized. Finally, she noticed that for the Simplex method to perform well, the initial vertices should define extreme ranges of the factors. 

Optimization methods calculate one best future state as optimal result. Mathematical algorithms e.g. SIMPLEX or Branch Bound are used to solve optimization problems. Optimization problems have a basic structure with an objective function H(X) to be maximized or minimized varying the decision variable vector X with X subject to a set of defined constraints 0 leading to max(min)//(X),Xe 0 (Tekin/Sabuncuoglu 2004, p. 1067). Optimization can be classified by a set of characteristics ... 

The movements of a Simplex in the direction of the optimum are defined by a number of logical mles, which are applicable to systems with any number of factors. The rules and calculations for moving the Simplex can be found elsewhere... 

The function is calculated in a user routine starting at line 900. On the input you should define the N+l vertices of the simplex. If you do not have a better idea, these can be generated by perturbing the elements of an initial guess ane-by-one. 

Figure 3.6. A Simplex optimization of a two-factor system. Numbers give the order of experiments. The first three experiments (points 1, 2, and 3) define the initial Simplex, and dashed-line arrows indicate which points are dropped in favor of new factor values.

A simplex is a geometrical figure defined by + 1 vertices in n dimensional space (e.g. in two dimensions this would be a triangle). The procedure is to calculate the energy at n + 1 points and then to replace the worst point (the point of maximum energy) by another P which is the other side of the hyperplane defined by the remaining n points. If P is the centroid of these n points, then the new point is obtained by invertingthrough that is... 

Let us suppose that we are interested in implementing this procedure in our laboratory and we fix the time of agitation at 10 min. So, we want to look for the RC (Xi) and pH (X2) values that provide the largest percentage lead recovery (Y), and we will use the simplex method defined by Spendley etal [16]. 

Now we calculate the coordinates of the two vertices that define the first simplex. Several ways have been proposed to do this, the simplest being the so-called initial corner simplex", which in general can be formulated as shown in Table 2.22. 

To compare this approach with the original simplex method, we will consider again the first initial simplex that we defined initially (see Table 2.29). 

We see that between R and W, the lead recovery (response) has improved to 71-50% =21%, so R should be the new B of simplex number 2. This suggests us that, perhaps, in the (W - R) direction we might obtain better responses. To assess this, a new point E (expanded) can be defined, whose coordinates are calculated as... 

Now, to define the following simplex (simplex number 2), we replace vertex W of the previous simplex (simplex number 1) by the expanded vertex E. We obtain simplex number 2 (Table 2.31), the size of which is larger than the previous one and, more importantly, it is oriented in a good direction of improvement. Figure 2.14 displays the effect of the expansion movement. 

The constrained optimization procedure, originally developed from the simplex method and first described by Box, is ideally suited to model refinement (.8). It is a search method that searches for the minimum of a multidimensional function within given intervals. It possesses all the advantages of search methods, among them that calculation of derivatives is not necessary, a test to assure the independence of variables can be omitted, and diverse variables can be easily included. These are exactly the requirements of model refinement where bond lengths, bond angles, torsion angles, and other parameters are used within experimentally defined limits. 

Assume we have to do optimization of a phenomenon that is defined by four factors (k=4). We use Table 2.209 to apply simplex optimization and define the initial simplex. To determine the operational matrix, we should know factor values in the center of experiment and their variation intervals. These data are to be found in Table 2.210. Real factor values are obtained from relation (2.59) ... 

By formation of the operational matrix of the initial simplex, we end up with the first problem of simplex optimization. One should now determine the direction of movement of the initial simplex to optimum. As said before, simplex movement is linked with rejecting the most inconvenient vertex and with defining conditions for... 

This is why, to find out the optimum, the simplex method has been applied once again, as shown in Table 2.247. It should be noted that simplex optimization trials are not now realized experimentally but that they are abstract ones where response values in simplex vertices are calculated from the obtained second-order regression model. It is evident that in this case the simplex method has been used to test the response surface that is defined by the obtained regression model. Trials 18-21 have given the best calculated response values, which is why they have been verified experimentally, Table 2.248. 

---