Study On The Improvement Of Multi-center Location Based On Max-min Distance Assembling Method

How to select logistics distribution center is a critical issue, it impacts the operation efficiency of whole supply chain it involved. Center of gravity method applies to many logistics distribution center selection due to it has the advantage of easy calculation and availability to select multiple continuous points. Compared with previous inspired algorithm, Center of gravity method can saving room during calculation and avoid trouble of multiple dimension, further avoid suffers in area optimal points.This paper introduced an improved center of gravity method, which is based on Max-min distance assembling Method, K-means algorithm, and isolated point filter algorithm based on density. We can name it as multiple distance Method based on max-min assembling algorithmFirstly, this paper not only applies assembling algorithm from data digging area to logistics distribution center selection, but also combines center of gravity method. The benefit is the new method can set the number of assembling points automatically, which improves algorithm efficiency and reduces overall cost. This method includes three phases: assembling area division, sites selection and cost analysis.Then, considering the disadvantage of traditional center of gravity method, it only consider transportation cost, and exclude some fixed cost and variable cost. I.e. construction cost and pricing difference and other factors. This paper makes and improvement to consider impacted factor as many as possible to improve. Firstly we set ground price with certain weight and put it at the beginning of calculation, then we got the result of optimal solution for the first round, and next step, involve fixed cost and operation cost as key factors to get the final result. Thus can avoid bad result, which is the more multiple assembling points, the lower the overall cost. this is not consistent with actual, so the final result make more sense.Next, this paper introduces the classic K-means assembling algorithm, administrative levels assembling algorithm, DBSCAN algorithm, assess their advantage and disadvantage, based on the comparison, bring a three partial assembling algorithm based on max and min method. Firstly, we set the number of assembling points based on max-min algorithm automatically, which is the advantage of this algorithm. Next step, classification based on K-means algorithm finally add filter mechanism in the 3rd part to avoid the impact of isolated points to the whole algorithm. This can not only improve assembling efficiency, improve robustness, but also avoid bad impact from isolated points to the final result.At last, to test its availability and efficiency when divide areas based on max-min three party algorithm. The paper does the simulation test to the algorithm listed below: three party method, administrative levels method, K-means method and DBSCAN method.From the test result, the solution based on max-min assembling algorithm is much better administrative levels method and K-means method. The result is quit close to simple K-means method. But it show advantage on efficiency and stability. Multiple selection method based on max-min assembling algorithm can not only divide assembling area with high efficiency, but also can reduce overall logistics cost, which can be applied to logistics distribution center selection with advantage.