SUMMARY
The collaborative planning in Supply Chain Management, within the context of uncertainty needs new systems that
minimize randomness. Uncertainty can be handled on the basis of heuristic algorithms that improve the logistic activities,
which include the Master Planning in Supply Chains, obtaining better results with regard to profit, cost minimization and
other targets that are pursued in the supply chain. This is for obtaining collaboration between the different
stages (supplier, manufacture, distribution and retailer) that the supply chain includes. In this approach, this thesis
presents a methodological proposal, which consists of a proposed model, description of the problem, two mathematic models
(deterministic and one of uncertainty), implementation of the model, the architecture and technology platform SCANN
(Supply Chain Artificial Neuronal Networks), and finally the application of this model and problem-solving tools for a
company. The two mathematical optimization models (Deterministic Mathematical Model “DMM” and Uncertainty Mathematical
Model UMM) consider simultaneously the profit maximization for comparing them. One model is set up in the deterministic
context and the other one uses the same, but applying uncertainty, which can be predicted by neural networks
(the neuronal networks are able to create data with a high quality using its own intern process and training data).
The models DMM, UMM and Neuronal networks are used on a technological Platform SCANN (developed by the doctorate),
which is applied in the ceramic sector that can be used in different models of the supply chain.
The technological platform SCANN can be used to take decisions in a centralized chain to a tactical-operative level.
The alternative of the decision, are defined by a decision-maker (taking into account his experience and knowledge),
this means, that he takes the decisions of the supply chain; all this taking into account the company’s politics.
This decision maker or can be unipersonal, collegiate, assigned, etc. The platform has been developed in VISUAL.NET and
this interacts with a mathematical determinist model which is developed in Mathematical Programming Language (MPL)
and with Artificial Neuronal Networks (ANN), on the possibility theory context.