Applying a Hybrid Algorithm to Solve Product Mix of Customer Orders Schedule Problems
關鍵詞：產品組合、顧客訂單、限制理論、粒子群演算法、變動鄰域搜尋法 ； Keywords: Product mix ； Customer order ； Theory of Constraints ； Particle Swarm Optimization ； Variable neighborhood search.
|Volume or Term/Year and Month of Publication||
|Academic Degree Category||
The study focused on online shopping companies with diversified customers’ orders under limited production resource environment. The objective of this research was to help companies make quick decisions of determining the optimal product mix of customers’ orders in order to maximize the total profits. Traditional customer orders problem considered that a single customer order had one product. However, this research considered that a single customer order contained multiple products, and considered the variability in product mix to satisfy customers’ orders. The study proposed a hybrid algorithm that combined particle swarm optimization algorithm (PSO) with variable neighborhood search (VNS), and used the bottleneck resource concept based on theory of constraints (TOC) to solve product mix of customer orders scheduling problems. Two neighborhood searching heurisitics were developed for the proposed VNS: the first one was to randomly select two confirmed orders and two unconfirmed orders in a feasible solution and switch them. The second one was to choose the lowest-profit confirmed order in a feasible solution, and randomly choose one unconfirmed order to switch them. This study used bit conversion (from the 10 numerical system to the 2 numerical system) to write programming codes, so that the particle velocity and location of customer orders scheduling problems would converge much fast. According to numerical examples, the results of this research showed that solution quality of the proposed hybrid algorithm was better than that of the basic PSO in four scenarios. Although the proposed hybrid algorithm required longer computation time, the solving time was within acceptable range.
工程學 > 工程學總論