Simplex designs

Spendley W, Next G R and Himsworth F R 1962 Sequential application of simplex designs in optimization and evolutionary operation Technometrics 4 441... 

Spendley, W., Hext, G.R., Hemsworth, F.M. Sequential Application of Simplex Design in Optimization and Evolutionary Operation, Technometrics, Njv. 1962, p. 441. 

Spendley, W. G. R. Hext and F. R. Himsworth. Sequential Application of Simplex Designs in Optimization and Evolutionary Operations. Technometrics 4 441-461 (1962). 

Figure 14.7 Optimization designs, (a) Centred composite design with three factors, (b) Three factors optimized in four steps using a simplex design.

Finally, the simplex design has also been adopted for crystallization purposes (Prater et ah, 1999). This is an iterative approach starting with one more combination than factors under investigation. In an example with three factors at three equally spaced levels, 0, p and q, the first set consists of... 

The Simplex Designs Straight lines in one dimension, triangles in two dimensions, tetrahedrons in three dimensions, etc. 

In some multicomponent mixture problems, it may be desirable to hold one component at some constant level and vary the others. In this case, we ignore the constant component and use the Simplex designs in the other components, assuming the constant total of these as 100% of the mixture. In the case of an oil additive, for example, we may wish to specify a mixture of additives which add up to 10% of the total composition. A possible design for a study of this type is given below ... 

One of the limitations to the Simplex design in additive experimentation is thet the desired total additive level is not always known. For example, we may be interested in obtaining a certain effect on a base stock using a combination of three additives. For cost reasons, we may limit the total concentration of additives to 6%. If we do this and fail to get the effect, we heve no information to estimate how much more additive we would need to get it. On the other hand, if we do get the effect at 6%, how do we know that we would not have obtained it less expensively at 4%. In situations of this kind, a central composite design can often be used, sat up in... 

Simplex design an experimental design based on the minimum symmetrical geometric pattern possible. This pattern will have a dimension of one less than the number of variables. Hence, for 4 variables, the Simplex design is a tetrahedron, for 3 variables a triangle, for 2 variables a line, and for a single variable a point. 

This example shows a design set up to determine the thermal stability of a mixture containing 3+85% of a total of three additives plus a constant amount (1 15%) of a fourth. The design is set up as a Simplex design in the three variable additives. Note in no case was any ingredient present alone the main area of interest in ingredient C was 0+35 to 1 25%,... 

Latin square designs, 1,52 simplex designs, 56,57,58 Experimental designs for specific problems, 61,62,63... 

Two consecutive contractions reduce the size of the simplex design. ... 

Optimization techniques may be classified as parametric statistical methods and nonparametric search methods. Parametric statistical methods, usually employed for optimization, are full factorial designs, half factorial designs, simplex designs, and Lagrangian multiple regression analysis [21]. Parametric methods are best suited for formula optimization in the early stages of product development. Constraint analysis, described previously, is used to simplify the testing protocol and the analysis of experimental results. 

The method of simplex design of experiments in mathematical theory of experiments belongs to a group of nongradient optimization techniques in multidimensional factor space. As a difference to gradient methods, this method does not require a mathematical model of researched phenomenon or does not require derivative of a response. 

Movement to optimum by a simplex method is done step by step and by comparing obtained response values, whereby a single response value is determined in each step. These properties of simplex design are considered its important advantage ... 

Simplex designs are applied in extreme experiments (min/max experiments) with a large reproducibility error-process noise ... 

Simplex designing is successfully applied for determining optimum of a research subject with more responses-multiresponse optimization ... 

Simplex designs of experiments were first published in 1962 [53], and since then application of this methodology has been constantly growing [31, 54, 10]. Under simplex designing, we understand finding of the optimum of a response function by moving simplex figure on the response surface. Simplex movement is done step by step, whereby in each new step-trial the simplex vertex with the most inconvenient response value is rejected. 

Factor-variation intervals in simplex designing are as a rule chosen so that they are five to ten times greater than the error square mean of their measurements. 

Table 2.208 offers data for forming initial simplex (k=3). All the data are divided by maximal value 0.612 and so we obtain a simplex design matrix. Prior to forming operational matrix, we use expression (2.59) so that ... 

The radiation dose should not be over the value of 3.5+4 Mrd. The simplex optimization method was selected since response and single factor limitations do not interfere with this method. Simplex design matrix is obtained from Table 2.208 for k=2. Matrix elements are divided by the highest value of 0.578 so as to get initial simplex matrix, Table 2.218. Formulas for transformation from coded to real factor values have these forms ... 

The term Simplex design (or Simplex lattice design) is unfortunate. It creates confusion between the Simplex method for optimization (section 5.3) and the Sentinel method of Glajch et al., which is a very different method by all accounts. To avoid further confusion, we will not use the word Simplex in connection with the Sentinel method. 

A total of seventeen different mobile phase compositions are required to generate the response surface for the two Simplex designs as shown in Figure 6, Each run is done in duplicate, and the entire design is randomized to reduce experimental bias. Runs 12 through 17 are the checkpoints used to determine variance and goodness of fit. The overall design is shown in Table II,... 

When the experimental region is not a priori known or when one is not interested in modelling the response but only in finding the optimal conditions, the simplex methodology may offer an alternative. This sequential method should not be confused with the simplex designs described in Section 6.5, which are mixture designs. 

Simplex designs are quite rarely used beeause sueh circumstances in which there are no composition... 

The simplex design for four variables shown in Table 11.4 was used to define the starting simplex. From Table 11.2 is seen that with four variables, the corresponding values of p and q are ... 

For new synthetic methods, it would be most helpful for future users, if the inventors include a sensitivity test of the suggested experimental conditions, e.g. by a simplex design, when the method is published. 

---