This last one is associated to a coverage of 100%% for both criteria, but if desired, the user can select the option with best ED value, in this case, Solution 17, with 21 products, score of 98.678 % and a pairwise coverage of 98.907 %.
In H1, we consider values ranging from 0% up to about 100%, which means that in the least accurate case, a solution with a cost twice as high as its lower bound will already be accepted.
It is found that the SI-SRINR and the non-SI-SRINR methods can achieve same optimal worst-case SR solution with almost 400 and 700 iterations, respectively.
The approach tested in this work consists in adapting the Pitzer model, initially designed for aqueous electrolyte solutions, to the case of solutions with a mixed solvent.
From looking at the mentioned clusters, one sees that this is not the case, the solutions with the highest MIA have all comparatively low TAA (clusters I and II) and the ones that have a high TAA only achieve an intermediate MIA (cluster III).
Furthermore, using an appropriate case of pore solution with diffusivities of 8.2 × 10−12 m2/s in LD C-S-H and 2.0 × 10−13 m2/s in HD C-S-H, estimated chloride diffusivities in cement pastes agree well with measured values, which can indirectly indicate this modelling work is reasonable.
In the case of buffer solution with pH = 7, silicon oxide surface charge will be charged negatively.
In other cases, stirring the solution with vigorous moving of small steel balls (~0.4 mm in diameter; Super micro-ball generously provided by Toyo Seiko Co., Japan) by magnetic force was also applied, of which effect was confirmed by diffusion of dye.
Numerical results show that the method is able to solve the problem efficiently for single and multiple load cases obtaining solutions with smooth boundaries.
In most cases, optimal solutions with prescribed terminal conditions can be obtained within a few seconds (without parallel calculations).
