Aditia Ginantaka

Traceability is the ability to verify the history
and location of a food product, thus providing information
on each supply-chain actor, who the immediate supplier is,
and to whom the product was sent. The information system
approach has been used to manage and integrate all such
information by collecting, storing, then retrieving data and
information about the product from earlier stages of the
production process. Besides documentation and information
sharing, such traceability information systems can also
support timely resolution of customer complaints. This
paper presents modelling of routing and handling time
prediction using a fuzzy associative memory (FAM)
method. As a first response to customers, information about
the time required to resolve an issue can be provided after
the source of a product defect has been traced. With
regards to handling, traceability can assist with several
issues, e.g., product replacement, product recalls based on
retrieval of the contact numbers of affected customers on a
recall list , and inspections at each production unit to
ensure food safety standards. Based on such activities, it is
assumed that the handling time will be affected by the size
of the product inventory that can be used to replace a
defective product, the amount of product that must be
recalled from the market, and the time required internally
for the inspection process, which is set as the FAM input
variable. A FAM is a set of fuzzy-set pairs (A, B) that maps an input vector fuzzy set A to an output vector fuzzy set
B. Our experiments show that, from such a FAM formulation,
one can obtain 27 rules. The FAM will encode a
fuzzy-set pair (A, B) to obtain matrix memories, denoted by
M. As the prediction result, the matrix B can be obtained
from the computational matrix A and the matrix M; For
instance, in case of a contamination incident in a fish
product with inventory conditions of as much as 4 tonnes,
the product recall amounts to 21 tonnes and the inspection
will take 25 h, while the results of the computational
experiment show that the total handling time for this case
will be 66 h with low error rates.