Nowadays, material planning and control strategies are becoming continuously complex tasks spanning from individual plants to logistic networks. In fact, this is the consequence of increasing intricacy in product variants and their respective convolution in networks' structures. Customers ask for specific products with individual characteristics that force companies for more clever performances by more flexibility. For doing so, the existing planning and control systems, which work based on central monitoring and controlling, show some limitations for organizing every operation on time or in the right time. Therefore, in the recent decade, a great attention is put on decentralized control and, to some extent, autonomy. This paper tries to investigate the possibility of combining this new research paradigm with existing strategies in production logistics, in order to improve material handling and control task according to material flow criteria. To show this, an exemplary plant after decoupling point out of a logistic network is considered for simulation and analysis. This combines Conwip system with learning autonomous pallets' concept in a discrete event simulation model. Several decentralized control scenarios are experimented and compared together. Here, the learn methodology is brought to pallets based on fuzzy rules and advantage of closed loop systems.