Simulation-based optimisation approach to multi-choice transportation problem

The classical transportation problem minimises the total costs of transportation of a unique product from various supply points (or warehouses) to demand points. The problem assumes that freight costs from source to destination are constant and that the supply and demand quantities are equal and strictly known, so the market for the product is well-balanced. It thus involves a special type of linear integer programming, which becomes stochastic since the constraints or parameters are random variables from a known or unknown distribution. Several studies have formulated well-known deterministic models under probabilistic restrictions. The transformed models mostly keep the confidence level at a given minimum constant or else minimise the error level. Also, there is a multi-choice stochastic transportation problem, which introduces several unit costs. In this study, we try to simulate Roy's (2014) multi-choice stochastic transport model with random supply and demand quantities from a given Weibull distribution and compare the results of distribution and total costs. As a result of the simulation, total cost value was estimated lower than the result of the problem.