An intelligent computational approach to the optimization of inventory policies for single company

This study develops and tests a computational approach for determining optimal inventory policies for single company. The computational approach generally comprises of two major components: a meta-heuristic optimizer and an event-driven inventory evaluation module. Meta-heuristic is a powerful search technique, under the intelligent computational paradigm. The approach is capable of determining optimal inventory policy under various demand patterns regardless their distribution for a variety of inventory items. Two prototypes of perishability are considered: (1) sudden deaths due to disasters and (2) outdating due to expirations. Since every theoretical model is specially designed for a certain type of inventory problem while the real world inventory problems are numerous, it is desirable for the newly proposed computational approach to cover as many inventory problems/models as possible. In a way, the proposed meta-heuristic based approach unifies many theoretical models into one and beyond. Experimental results showed that the proposed approach provides comparable results to the theoretical model when demand follows their assumption. For demands not well conformed to the assumption, the proposed approaches are able to handle it but the theoretical approaches do not. This makes the proposed computational approach advantageous in that it can handle various types of real world demand data without the need to derive new models. The main motivation for this work is to bridge the gap between theory and practice so as to deliver a user-friendly and flexible computational approach for rationalizing the inventory control system for single company