existence is inevitable in
companies of this type, although including them also complicates the
471
search for useful solutions.
472
Accordingly, the search for a near optimal strategic plan was tested under the following three
473
theoretical scenarios involving optimization constraints:
474
5.2.1
Unrestricted production
475
First, the developed methodology was tested in a scenario without any operational or commercial
476
constraints. This enables the proper functioning of the methodology to be tested in a situation in
477
which every candidate solution within the search space constitutes a valid alternative. This
478
facilitates the process and means that the number of particles and interactions can be lower. In
479
this
scenario, the five parameters of the PSO algorithm were as follows: the population was 90
480
particles with a maximum number of iterations of 30, while the inertia,
cognitive and social
481
components each took the value of 0.5. Appropriate parameters selection is a fundamental aspect
482
of PSO and it is discussed further in Section 5.
483
Table 4 shows how, when no constraints force the different cages to
adapt to each other, all the
484
three cages tend to choose the same strategy, which we assume to be optimal: harvesting in the
485
same week and taking the same feeding decision. Together with the practicality of the
selected
486
strategies, as we will see later, this suggests the proper functioning of the methodology right from
487
the start. As
an exception, it is possible to see small differences in some points that would
488
undoubtedly be solved with more computing time.
489
Do'stlaringiz bilan baham: