A hybrid heuristic for the inventory routing problem under dynamic regional pricing

The inventory routing problem (IRP) seeks to meet the demands of customers during consecutive time periods. Because of the geographical distribution of customers and variations in willingness to pay of the consumers in distinct locations and time, regional and time-based pricing are powerful ways to improve profitability. In this study, a quadratic mixed-integer programming model for single product, multi-period Inventory Routing under the dynamic regional pricing problem (IRDRP) has been proposed. A hybrid heuristic approach is developed to solve it. This algorithm comprises five phases: initialization, demand generation, demand adjustment, inventory routing, and neighborhood search, which are embedded in a simulated annealing framework. Experimental results indicate as the problem size increases, the difference between CPLEX and the proposed heuristic algorithm optimality gap exhibits an upward trend and that the heuristic outperforms CPLEX. A sensitivity analysis demonstrates that by intensifying the scarce capacity, approaching an optimal solution will be more difficult.