Impact of displayed inventory level in a two-warehouse system for deteriorating item with discount facilities via different metaheuristic algorithms.

Improving the balance between exploration and exploitation stages is very crucial for preventing local optima when using metaheuristic algorithms to solve optimization problems. This is made by a number of procedures, including hybridization and search agent improvement. This work deals with the modelling of single deteriorating item based two-warehouse inventory system with variable demand depending on time, frequency of advertisement and displayed inventory level. Customers’ waiting time dependent partially backlogged shortages are allowed. Moreover, price break facilities are available for the retailer. In the proposed model, transportation costs for replenishing the items as well as for transferring the item from rented warehouse to owned warehouse are considered. Based on displayed inventory level and capacity of owned warehouse, different cases are investigated and the corresponding optimization problems are developed. An extended version of hybrid Tournament Differential Algorithm (TDE) is applied to solve the said optimization problem. Then the same problem is solved by differential evolution and teaching learning based optimization algorithm and the obtained results are compared with TDE algorithm. Finally, a numerical example along with sensitivity analysis (graphically) is carried out to illustrate the proposed model and also to explore how the changes in key model parameters effect on optimal policy. To analyze the statistical interpretation of different metaheuristic algorithms, a non-parametric test is performed. [ABSTRACT FROM AUTHOR]