Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
doi:10.21311/001.39.4.18 Modeling and Optimizing Location Selection of Urban-rural Logistics Center Based on SCP-PCA under E-commerce Environment
Lijuan Huang, Wende Zhao, Yue Wang School of Business Administration, Guangzhou University, Guangzhou, 510000, China
Raoyi Huang* Affiliated Middle School, Guangzhou University, Guangzhou, 510000, China *Corresponding author (E-mail: [email protected])
Abstract There is a largest income gap between China's urban and rural areas in the world. With the rapid development of China's e-commerce, scientific and reasonable logistics center location planning can help to narrow the gap and promote the integration of urban and rural development. This paper firstly analyzes the current study of logistics center location combined with domestic and overseas literature. Then according to Set Covering Problem-model (SCP) to get the initial location of logistics center, and make further evaluation and optimization based on Principal Component Analysis-model (PCA). Finally, in view of Guangdong Province is the largest one of urban-rural economic disparity in China, the paper takes Guangdong Province as an example to carry out an empirical study of logistics center location selection. The results have been proved to be robust and reliable, which can provide advice and support for making scientific policy of logistics development between urban and rural areas to government at different levels.
Key words: Logistics Center, SCP-PCA, Location Optimization, Urban-rural Integration.
1. INTRODUCTION
The idea, urban-rural integration, was first proposed by the west urbanists Ebenezer (1898) in the book of Tomorrow: A Peaceful Path to Reform. The urban-rural integration under e-commerce environment (abbr. UIEE) often refers to use e-commerce theory and technology to achieve integration of urban and rural areas or narrow the gap between urban and rural areas. In this paper, urban-rural integration mainly focused on the logistics integration of urban and rural areas based on electronic business, promoting economic development of urban and rural areas via optimizing the logistics center location, integrating urban and rural logistics resources, combining with electronic business. There are many different approaches to discuss logistics center location from domestic and oversea in recent years. The location selection of distribution center is one of the most important decision issues for logistics managers, and the concepts of fuzzy numbers and linguistic variables can be applied to evaluate both the linguistic and numerical data without any constraints(Chen, 2001). Fuzzy group decision making based on extension of TOPSIS method has been proposed for facility location problem, the decision criteria are favorable labor climate, proximity to markets, community considerations, quality of life, proximity to suppliers and resources (Ertuğrul, 2011). A new hybrid heuristic algorithm which combines rough set methods and fuzzy logic can be feasibly applied to solve other industrial decision-making problems, such as facility layout problems, suppliers selection problems, etc.(Chanet al., 2011). A modified particle swarm optimization for disaster relief logistics under uncertain environment which is formulated as a mixed-integer nonlinear programming can minimize the sum of the expected total cost and the variance of the total cost (Bozorgi-Amiri and Jabalameli, 2012).Trends relating to the increasing size of cities and the growing number of elderly persons with healthcare needs are presenting substantial treats to liveability and mobility. Designing and evaluating of City logistics become more and more significant (Taniguchia and Thompsonb et al., 2014). City Logistics Centers (CLC) are an important part of the modern urban logistics system, and the selection of the location of a CLC has become a key problem in logistics and supply chain management(Goh and Rao, 2015).The multi- objective optimization model for the cold chain distribution center location has important theoretic value and practical significance for the optimization of cold chain distribution system, which can provide reference for the cold chain distribution enterprises(Zhuand Hu, 2012). Location strategy of logistics distribution center based on hierarchical genetic algorithm has a positive significance to optimize the structure of logistics system and improve the efficiency of logistics system(Liand Du etal., 2012). The research results are in accordance with the reality, which has proved that the logistics center location evaluation system of county-level city for
133
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016 domestic logistics companies is valid and operative(Yu and Chen, 2015). Logistics center plays a very important role in the logistics system operation, it is a bridge connects logistics upstream supplier and downstream demand (Wang and Zhang, 2015). According to the domestic and overseas literatures in recent years, we summarize that most scholars studied logistics center location from the perspective of the city or rural. First determined the alternate address of logistics center by continuous or discrete locating, then evaluated by Analytical Hierarchy Process (AHP) or fuzzy comprehensive evaluation method. Although this approach offers a certain reference value, it cannot play an objective and correct evaluation role due to the complexity of influencing factors. Therefore, there are two innovations in this paper: first, we study the logistics center location from the new sight of urban-rural integration under e-commerce environment (UIEE) instead of the traditional single perspective of city or rural; second, according to the principles of the location of logistics center, we achieve logistics center location optimization based on Set Covering Problem-model and Principal Component Analysis-model (SCP-PCA).
2. BASIC WORKING PRINCIPLE
2.1. Modeling Procedure Modeling Procedure of SCP-PCA is showed in Figure1, which is an optimization system with feedback loop. At first, determining the research object of this paper is that logistics center location optimization based on urban-rural integration; then, setting the overall goal according to the research object, and we should follow the principle of adaptability, economy, coordination and strategy; at last, proposing the optimized logistics center location model based on SCP-PCA. After filtering the data sets, determining the alternate address of logistics center based on SCP and Lingo, and judging whether it meets the quantity optimal principle of logistics center or not, otherwise it will be re-solved. Then, constructing the evaluation index system and optimization model based on PCA, and judging whether the result conforms to the principle of logistics center location. Finally, optimizing the above results. There are two key steps of the model. (1) We determine the alternate location of logistics center based on SCP. (2) The result of logistics center location is optimized based on SCP-PCA. The two steps are described as follows.
Data Location Object Target setting Principle SCP Model Model Solution matrix
Yes If not conform to it Selecting data sets No Location Covering
optimiz Yes Evaluation Data Evaluation PCA model Yes ation results collection system
Figure 1. The procedure of modeling
2.2. Model Design Model Design is based on the SCP, which is a classical model in computer science and complexity theory. SCP is one of most important discrete optimization model because of its practicability. Real world problems that can be modeled as set covering problem include facility location problem, airline crew scheduling, nurse scheduling problem, resource allocation, assembly line balancing, vehicle routing, etc. SCP is also a problem of covering the rows of an m-row/n-column zero-one matrix with a subset of columns at minimal cost (Gouwanda and Ponnambalam, 2008). SCP can be formulated as follows:
Minimise z j (1) jN
(2) subject to aij z j 1, j 1, , m jN (3) Zj （0,1） j 1, , n
Equation (1) is the objective function of set covering problem, where Zjis decision variable. Equation (2) is a constraint to ensure that each row is covered by at least one column where aij is constraint coefficient matrix of
134
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016 size m×n whose elements comprise of either „1‟ or „0‟.Lastly,equation (3) is the integrality constraint in which the value is represented as in Equation (4).
1 if j S (4) z j 0 otherwise Even though it may seem to be a simple problem by judging from the objective functions and constraints of the problem, SCP is a combinational optimization problem. It has been proven to be NP-Complete decision problem.
2.3. Model Optimization Although the alternate location of logistics center is set up based on SCP, the factors of logistics center location involves not only transportation and construction costs(Marco and Caglianoetal., 2014). We should also consider the factors such as economy, society, logistics, e-commerce etc.(Zak and Weglinski, 2014). PCA is a traditional multivariate statistical method commonly used to reduce the number of predictive variables. PCA looks for a few linear combinations of the variables that can be used to summarize the data without losing too much information in the process. The details are shown as follows(Hinton and McMurray,etal., 2014). (1)The standardization of the original index data. Acquiring p-dimensional randomvector
XXXXi (,,,) i12 i ip , in（1,2, , ）, np , Structuring sample matrix and standardizing it as follows:
xxij j Eij (5) s j n n x ()xx 2 Wherei(1,2, , n ); j (1,2, , p ) , and x i1 ij , S 2 i1 ij j . j n j n 1 (2)Calculation of correlation coefficient matrix R. EET R[] r xp (6) ij p n 1 EE Where r kj kj , i,jp 1,2,.... . ij n 1
(3)Solving characteristic equation，let RI-0 p , and get characteristic root. To determine the value of
m j1 j mby p 0.85 , and the utilization rate of information is more than 85%. Solving equations Rb j b and j1 j 0 get unit feature vector bj , where j =1, 2…m.
T 0 (4)Transforming the standard index variables into some main components.Uij Eij b, j 1,2, , m , and
U1is named for the first principal component, U2is named for the second principal component, Up is named for the principal component p. (5) Comprehensive evaluation. The principal component is weighted sum to obtain the final evaluation value, and the weights are contribution rate of principal components.
3. THEEMPIRICALANALYSIS
Guangdong, as the most important provinces for China's economy, is also one of the largest provinces of wealth gap between urban and rural areas. The lack of effective integration of logistics resources is the most urgent problem. This paper made an empirical analysis of Guangdong province.
3.1. Preliminary Planning in Guangdong Province
(1)Key Parameters According to the model of SCP, we need to count the distance between each alternative logistics center. In terms of transportation, distance is proportional to time, so we can consider the data calculated by distance
135
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
function as time (unit: day). There are 21 cities in Guangdong province, and each city may be considered as an alternative logistics center, the specific distance is shown in Table 1 and the code from 1 to 21 reference to the cities of Guangdong province. The details are shown in Table 2. Based on the requirements of logistics center location, we generally choose the city which has large logistics volume as logistics center, this will reduce the logistics cost and time. In order to save the construction and transportation cost of logistics centers, we need to select as few city as possible to build logistics centers. Economic mileage of road transportation is from 300km to 500km, it means that goods need to be delivered in 1.5 days once ordered. Concerning about timeliness and special attributes of the rural logistics, in this paper, we set economic mileage of road transportation from 100km to 200km. After modeling and debugging, we finally determined 150km as economic mileage of road transportation and labored it later. Specific analysis is as follows.
Table 1. Distance matrix City 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 1 139.2 128 442.2 34.8 222.1 419.5 111 95.3 340.7 141.5 384.2 274.1 199.5 228.3 79 71.9 87.4 460.5 431.1 145.9 2 139.2 154.8 339.2 141.8 339.7 496.7 237.1 150.6 417.9 90.8 371.9 171.1 187.3 305.6 200.4 74.6 126.2 355.5 328.2 272 3 128 154.8 473.4 130.8 350.7 407.7 197.4 90.1 328.9 189.9 461.3 305.3 276.6 202.4 211 125 49.5 491.8 462.4 233 4 442.2 339.2 473.4 461.4 634.3 812.9 539.2 470.2 734.1 312.9 156.4 187.2 291.8 625.2 494.9 375.5 445.8 46.7 46.4 584.2 5 34.8 141.8 130.8 461.4 248.2 395.7 97.5 79.1 316.9 167.8 413.3 291.6 228.7 212.1 97.9 98.2 79.4 485.3 448.6 133.8 6 222.1 339.7 350.7 634.3 248.2 651.1 336.5 307.1 572.3 334.3 484 466.8 367.5 440.1 163.5 275.5 322.6 653.2 473.4 348.9 7 419.5 496.7 407.7 812.9 395.7 651.1 450.4 362.7 100.1 538.8 804.3 654.3 619.7 215.3 488.9 474 404.4 840.8 804.7 460.4 8 111 237.1 197.4 539.2 97.5 336.5 450.4 118.2 363.8 243.2 488.4 375.4 303.7 249 162.9 173.3 159.9 561.8 525.8 56 9 95.3 150.6 90.1 470.2 79.1 307.1 362.7 118.2 275.6 185.3 470.2 300.8 285.6 163.2 154.8 120.4 50.6 485.7 457.8 151.1 10 340.7 417.9 328.9 734.1 316.9 572.3 100.1 363.8 275.6 481.3 719.4 569.4 534.8 130.4 404 389.1 319.2 754.4 726.5 251 11 141.5 90.8 189.9 312.9 167.8 334.3 538.8 243.2 185.3 481.3 271.6 144.2 86.9 342.3 204.7 92.6 162.9 329.2 301.3 294 12 384.2 371.9 461.3 156.4 413.3 484 804.3 488.4 470.2 719.4 271.6 263.2 191 604.7 420.3 358.4 432.1 134.9 112.2 523.9 13 274.1 171.1 305.3 187.2 291.6 466.8 654.3 375.4 300.8 569.4 144.2 263.2 218 456.7 326.4 207 277.3 202.6 174.7 415.8 14 199.5 187.3 276.6 291.8 228.7 367.5 619.7 303.7 285.6 534.8 86.9 191 218 420.6 236.1 174.2 247.9 268.9 246.1 339.7 15 228.3 305.6 202.4 625.2 212.1 440.1 215.3 249 163.2 130.4 342.3 604.7 456.7 420.6 287.1 272.2 218.4 637.5 609.5 166.4 16 79 200.4 211 494.9 97.9 163.5 488.9 162.9 154.8 404 204.7 420.3 326.4 236.1 287.1 135.6 159.3 511.9 483.9 196.5 17 71.9 74.6 125 375.5 98.2 275.5 474 173.3 120.4 389.1 92.6 358.4 207 174.2 272.2 135.6 97.5 391.5 363.6 203.1 18 87.4 126.2 49.5 445.8 79.4 322.6 404.4 159.9 50.6 319.2 162.9 432.1 277.3 247.9 218.4 159.3 97.5 461.1 433.2 196 19 460.5 355.5 491.8 46.7 485.3 653.2 840.8 561.8 485.7 754.4 329.2 134.9 202.6 268.9 637.5 511.9 391.5 461.1 29.8 601.7 20 431.1 328.2 462.4 46.4 448.6 473.4 804.7 525.8 457.8 726.5 301.3 112.2 174.7 246.1 609.5 483.9 363.6 433.2 29.8 571.1 21 145.9 272 233 584.2 133.8 348.9 460.4 56 151.1 251 294 523.9 415.8 339.7 166.4 196.5 203.1 196 601.7 571.1
Table 2. City code Code 1 2 3 4 5 6 7 8 9 10 11 City Guangzhou Shenzhen Zhuhai Shantou Foshan Shaoguan Zhanjiang Zhaoqing Jiangmen Maoming Huizhou Code 12 13 14 15 16 17 18 19 20 21 City Meizhou Shanwei Heyuan Yangjiang Qingyuan Dongguan Zhongshan Chaozhou Jieyang Yunfu
(2) Preliminary results
Defining Zj=1 if column j is in the solution and Zj=0 otherwise, we set Zj= {Guangzhou, Shenzhen, Zhuhai,Shantou,Foshan,Shaoguan,Zhanjiang,Zhaoqing,Jiangmen,Maoming,Huizhou,Meizhou,Shanwei,Heyuan, Yangjiang,Qingyuan,Dongguan,Zhongshan,Chaozhou,Jieyang,Yunfu},andj={1,2,3,...,21}. After calculation, we select Foshan, Shaoguan, Maoming, Huizhou and Jieyang as logistics centers in Guangdong province, which can cover all demand points. The details are shown in Table 3. The logistics center location selection involves economy, society, transportation, infrastructure, hydrology, geography, etc. many factors. Therefore, the results of logistics center location belong to the initial planning. We need to use comprehensive evaluation method based on PCA for further optimization.
136
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
Table 3. Alternate location in 150 kilometers Alternate Location Covering Foshan Guangzhou Shenzhen Zhuhai Foshan Zhaoqing Jiangmen Qingyuan Dongguan Zhongshan Yunfu Shaoguan Shaoguan Maoming Zhanjiang Maoming Yangjiang Huizhou Guangzhou Shenzhen Huizhou Shanwei Heyuan Dongguan Jieyang Shantou Meizhou Chaozhou Jieyang
3.2. Evaluation and Optimization (1) Assessment index system We should follow the principle of comparability, integrity, scalability, non-overlap, qualitative and quantitative analysis for index selection(Nick T. Thomopoulos, 2016). And combining the perspective of urban and rural integration, the details of the evaluation index system are shown in Table 4. Data in this paper is from 2014 statistical yearbook of Guangdong Province. The values of evaluation index of logistics center in Guangdong province are shown in Table 5.
Table 4. Evaluation index system Consumption index of urban Urban economic index Agricultural economic index and rural areas
Gross output Rural per Yield of Purchases, sale GDP value of Total sown Total import and GDP capita net agricultural s and inventory Per Capita agricultural area of crops export commodities (100 million yuan) income products of wholesale (yuan) products (100 (MU) (100 million yuan) (yuan) (ton) (10000 yuan) million yuan)
X1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 E-commerce index Urban and rural logistics index Total The revenue E-commerce Third industry Total freight volume of investment in of Post and transaction added value (100 volume cargo turnover fixed assets telecommunic Mileage (km) volume million yuan)) (million tons) (100 million t) (100 million ations (100 (billion yuan) yuan) million yuan)
X 9 X10 X11 X12 X13 X14 X15
Table 5. Data of evaluation index
City X 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 X 9 X 10 X 11 X 12 X 13 X 14 X 15 1 15420.14 119695 11688.2 435548 389.98 1346234 275130739 7241 9963.9 4594 82052 6563.75 4447.3 568.74 9004 2 14500.23 136948 41548 59 13.95 155 120219098 32732 8198.14 4320 29226 2090.03 2490.2 598.1 1680 3 1662.38 104786 11415.7 41383 79.05 107013 26776016 3306 770.21 495 8457 133.39 960.89 54.39 1447 4 1565.9 28661 7944.5 452866 157.73 1069426 8505991 562 660.94 467 4628 183.93 780.9 75.39 3802 5 7010.17 96310 12693.9 98193 263.52 310139 38997507 3894 2530.76 2089 27206 240.72 2375.6 165.35 5204 6 1010.07 35063 7325.95 860558 214.12 2356163 2486514 141 450.51 301 12184 255.05 664.52 27.49 15273 7 680.33 22499 6946.14 886208 132.73 2456656 684749 197 260.15 203 3995 60.47 342.73 26.36 15346 8 800.01 18603 7353.32 1199269 269.52 3220412 1548543 107 345.73 238 6325 133.26 280.5 46.72 16961 9 2678.35 57144 9465.1 557041 217.77 1749876 7287487 3495 991.09 798 19063 336.3 1401.3 86.44 11234 10 671.75 22560 7806.82 417222 173.67 1429413 711989 254 247.81 200 1934 26.07 462.09 23.05 5470 11 5490.02 66109 17003.4 12435 33.15 41115 26698506 9322 2951.06 1636 12863 432.27 1383.9 254.73 5002 12 2638.93 83393 12311.7 72686 110 224046 13192497 2169 1108.35 786 16719 146.55 962.93 87.33 2589 13 2000.18 44546 9396.34 926671 286.48 2867304 8285657 1202 828.34 596 9999 135.07 1000.8 60.07 10012 137
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
14 1039.84 42017 9171.2 698220 307.95 2187802 833274 145 333.14 310 7672 152.64 598.66 28.09 7473 15 2060.01 28859 7906.6 1413023 640.53 4320514 10913858 336 824.24 614 10590 388.21 795.58 76.45 21800 16 2160.17 36063 7624.93 1421427 567.58 3731273 14342319 74 893.66 644 7123 162.06 660.53 53.53 15642 17 1660.07 41479 7502.95 1122736 385.71 3029678 4795644 427 606.59 495 4472 60.39 1007.8 41.13 13382 18 1093.04 28928 7350.7 759991 254.6 2688106 1605994 265 494.65 326 10363 179.57 505.97 36.86 21746 19 780.34 28837 8035.34 266392 90.72 663990 2178337 239 289.48 232 3944 175.63 253.63 27.26 5048 20 1605.35 26866 6594.58 832423 229.66 2053386 7432666 286 437.18 478 2784 51.18 829.39 45.36 7210 21 602.3 24863 7934.74 684649 210.35 1761599 2372775 96 207.42 179 4410 73.72 623.38 21.53 7588
Then data was analyzed with SPSS and we got results based on PCA. The main results are as follows. From Table 6, 9.741 as the eigenvalue of first principal component, which accounted for 64% variation of the original variable. 22.203% is the variation of the original variable that the second principal components can be explained, and 7.518% is for the third. Together the first three principal components can explain 94.660% of the original variable information.
Table 6.Total variance explained Table 7.Rotated component matrix Extraction Sums of Square Component Initial Eigenvalues Loadings Component 1 2 3 % of Cumulati % of Cumulati Total Total X 0.976 0.173 0.091 variance ve % variance ve % 1
1 9.741 64.939 64.939 9.741 64.939 64.939 X2 0.897 -0.154 -0.005 2 3.33 22.203 87.142 3.33 22.203 87.142 X3 0.75 -0.312 0.577 3 1.128 7.518 94.66 1.128 7.518 94.66 X4 -0.591 0.737 0.259 4 0.368 2.451 97.112 X5 -0.246 0.871 0.042 5 0.23 1.534 98.646 X6 -0.578 0.763 0.261 6 0.096 0.642 99.288 X7 0.89 0.373 -0.203 7 0.059 0.392 99.68 X8 0.797 -0.23 0.553 8 0.021 0.14 99.82 X9 0.968 0.222 0.052 9 0.013 0.085 99.905 X10 0.976 0.173 0.091 10 0.006 0.041 99.946 X 0.844 0.412 -0.3 11 0.006 0.038 99.984 11 12 0.001 0.009 99.993 X12 0.813 0.461 -0.259
13 0.001 0.005 99.998 X13 0.912 0.282 -0.198
14 0 0.002 100 X14 0.969 0.128 0.166
15 -4.26E-16 -2.84E-15 100 X15 -0.464 0.739 0.258
Table 8. Principal components Principal Index Load Component naming Component
X X X X X 1 , X 2 , 3 , 7 , 8 , 9 , The city's comprehensive economic strength Y X 0.75-0.976 1 X 10 , X 11 , X 12 , 13 , X 14 component
, , , 0.737, 0.871, 0.763, Agricultural economy and infrastructure Y X 4 X X X 2 5 6 15 0.739 component
Y X , X 8 0.577,0.553 The rural consumer ability to export component 3 3
138
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
Loading matrix of the first three factors is given in Table 7. From the matrix, the first three factors can be redefined as three new principal components, specifically details are shown in Table 8. According to the relationship between the component matrix and eigen value, the corresponding feature vectors of the three principal components can be calculated by using the component matrix and the eigen values in Table 6 and Table 7.
According to the formula (7),the score of Y1 、Y2 、Y3 can be calculated , the formula (7) is as follows: YCX * . ii (7)
X Y 1 C 0.313 0.287 0.149 1 X 1 WhereYY , X 2 CC0.095 -0.084 0.405 . Then according to the formula (8), i 2 i 2 Y3 C3 0.085 -0.005 0.243 X15 comprehensive score can be calculated.Formula (8) is shown below.
123 YYYY***1 2 3 (8) 1 2 3 1 2 3 1 2 3
And 1 , 2 and 3 are eigenvalues of three principal component, the concrete value are 0.686, 0.235, 0.079. Calculating principal component score and ranking in the following table.
Table 9. Comprehensive score and ranking of 21 cities in Guangdong Province
City Y1 Ranking Y2 Ranking Y3 Ranking Y Ranking
1 78196838.55 1 56991115.25 1 -52160121.87 21 62867751.33 1 2 34356849.04 2 24561438.13 2 -22944907.03 20 27507416.05 2 3 11087215.8 3 8139727.55 3 -7350275.99 19 8931245.2 3 4 7644802.82 4 5526354.85 4 -5075685.72 17 6137433.092 4 5 7635308.81 5 5476121.48 5 -5077587.21 18 6118987.909 5 6 3142862.04 7 5070189.63 6 -1471552.34 15 3228286.469 6 7 3740082.52 6 2815727.94 9 -2444930.01 16 3031948.506 7 8 2055794.17 9 4616280.31 7 -672695.87 11 2439531.2 8 9 2153865.03 8 2367956.06 13 -1248137.83 14 1933811.781 9 10 1674231.45 10 3268221.75 8 -647509.61 10 1863615.476 10 11 1591695.28 12 2714638.08 10 -708110.7 12 1672350.142 11 12 1673527.93 11 2449050.88 12 -822446.93 13 1657120.563 12 13 608055.11 13 2702687.5 11 108205.85 8 1059586.619 13 14 123452.98 16 1846125.98 15 317979.62 6 542914.0082 14 15 230380.43 15 1498993.3 18 151372.48 7 521623.4675 15 16 458887.04 14 829832.47 21 -183642.56 9 494838.5081 16 17 -173265.7 18 1767021.67 16 545524.42 4 338879.6761 17 18 -375038.24 20 2153208.33 14 794318.25 1 310791.4096 18 19 -283760.03 19 1367157.62 19 553375.71 3 169922.9678 19 20 -132073.05 17 911362.27 20 322157.67 5 148722.2079 20 21 -420050.48 21 1529315.19 17 695462.3 2 125740.4978 21 (2) Result and Analysis According to Table 3 and Table 9, there are five potential sites for logistics center location. All sites are top-ranking by comparison except Maoming and Shaoguan. And they are able to bear the cost of logistics center construction and the amount of logistics on the basis of meeting the principle of timeliness. In our study, we can further adjust and optimize it, the details are as follows: First, Guangzhou and Foshan have good economic
139
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016 foundation and their coverage are approximately equal. But in view of internationalization and long-term perspective, Foshan is replaced by Guangzhou as the rural logistics center, which is more appropriate. Then, Maoming‟s economy is relatively behind, it must bound to the construction and development of logistics center. However, Yangjiang is not far from the Maoming and the rank is eight, and its economic foundation is better, which will give full play to the role of logistics center in driving the development of surrounding rural economy. Therefore, Maoming is replaced by Yangjiang. According to the principal component score and ranking, we add Zhuhai as the new logistics center. Together with the five cities above, logistics center location in Guangdong Province is determined, they are Guangzhou, Huizhou, Zhuhai, Yangjiang, Jieyang and Shaoguan. The location of the logistics center in this paper is not the exact location but adjustment of the logistics center location area based on 150 km economic scope. The specific location and coverage are shown in Table 10 and Figure 2. This location planning is expected to save mileage of 1800 km and reduce logistics cost 1%-2%.
Table 10. Optimization ranking Ranking Alternate Location Covering 1 Guangzhou Shenzhen Foshan Zhaoqing Qingyuan Dongguan Yunfu 4 Zhuhai Jiangmen Zhongshan Zhuhai 5 Huizhou Shanwei Heyuan Huizhou 8 Yangjiang Maoming Yangjiang Zhanjiang 11 Jieyang Meizhou Chaozhou Jieyang Shantou 14 Shaoguan Shaoguan
Shaoguan
Qingyuan Meizhou Heyuan Chaozhou Zhaoqing Guangzhou Jieyang Huizhou Shantou Foshan Yunfu Shanwei Dongguan ZhongshanShenzhen Jiangmen Zhuhai Yangjiang Maoming
Zhanjiang
Figure2. Optimization map of location selection
4. CONCLUSIONS
In conclusion, to minimize the logistics cost and maximize the integration of urban-rural resources, scientific and reasonable logistics center location planning has become one of the most important issues for urban-rural coordinated development that has major impact on harmony in the long run. We construct an optimized model of logistics center location based on SCP-PCA in the perspective of UIEE and we apply the model to the case of Guangdong Province:
140
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
(1)This paper researches the optimization of logistics center location from the perspective of UIEE, abandoned the traditional single city perspective, which can reduce logistics cost and promote integration between urban and rural area. (2) We construct a closed-loop location optimization model based on SCP-PCA, rather than the traditional open-loop system, which makes optimization result more reasonable and scientific. (3)This model has realized the logistics center location optimization in Guangdong Province, and the application effect is remarkable. It can be considered that the model can be applied to other regions of the world where the urban-rural income gap is severe. Scientifically to plan and construct logistics center location based on SCP-PCA in the perspective of UIEE is a complicated systematic engineering. Due to the limited capacity and limited time of the authors, this paper has not good consideration for construction cost of logistics center. We will add more factors such as fixed cost of logistics center construction and variable cost of operation for making further improvement of the location model.
Acknowledgements The authors thank the anonymous reviewers for their valuable remarks and comments. This work is supported by 2015 National Social Science Fund of China (Grant No. 15BGL201), the key project of the National Social Sciences Program Fund (Grant No. 14ZDA031), the 12th Five-year Philosophy and Social Science Planning Project of Guangdong Province in 2014(Grant No. GD14CGL05), the Characteristic Innovation project of Ordinary Colleges and Universities in Guangdong in 2014 (Grant No. 2014WTSCX057), 2015 provincial Comprehensive Teaching Reform of Electronic Commerce in Guangdong, Cultivation Project of Teaching Achievement Award in Guangdong Province In 2014 (Exploration and practice on the training mode of e-commerce talent based on the IUR cooperative innovation), Key Project of Graduate Education Innovation Program of Guangdong Province (Grant No. 2015JGXM-ZD21), the 11th Five-year Social Science Planning Project of Jiangxi Province in 2010(Grant No.10GL35), and 2012 Teaching Reform Project of Colleges and Universities of Jiangxi Province(Grant No. JXJG-12-3-16).
REFERENCES
Ali Bozorgi-Amiri, Mohammad Saeid Jabalameli, Mehdi Alinaghian, Mahdi Heydari (2012) “A modified particle swarm optimization for disaster relief logistics under uncertain environment”, The International Journal of Advanced Manufacturing Technology, 60(4), pp.357-371. Alberto De Marco, Anna C. Cagliano, Giulio Mangano, Francesca Perfetti (2014) “Factor influencing Logistics Service Providers Efficiency‟ in Urban Distribution Systems”, Transportation Research Procedia, 3(10), pp.499-507. Chen Chen-Tung (2001) “A fuzzy approach to select the location of the distribution center”, Fuzzy Sets and Systems, 118(2), pp.65-73. Congjun Rao, Mark Goh, Yong Zhao, Junjun Zheng (2015) “Location selection of city logistics centers under sustainability”, Transportation Research Part D, 36(03), pp.29-44. Chang-bing Li, Mao-kang Du, Hui-ying Cao (2012) “Location strategy of logistics distribution centers based on hierarchical genetic algorithm”, Application Research of Computer, 29(01), pp.1-4. Darwin Gouwanda, S. G. Ponnambalam(2008)“Evolutionary Search Techniques to Solve Set Covering Problems”, World Academy of Science, Engineering and Technology, 39(01), pp.1-6. Eiichi Taniguchia, Russell G. Thompsonb, Tadashi Yamada (2014) “Recent Trends and Innovations in Modeling City Logistics”, Procedia-Social and Behavioral Sciences, 125(03), pp.4-14. İrfan Ertuğrul (2011) “Fuzzy Group Decision Making for the Selectionof Facility Location”, Group Decision and Negotiation, 20(11), pp.725-740. Perry R. Hinton, Isabella McMurray, Charlotte Brownlow (2014) SPSS Explained.Routledge: New York. Jacek ZAK, Szymon WEGLINSKI (2014) “The selection of the logistics center location based on MCDM/A methodology”, Transportation Research Procedia, 3(10), pp.555-564. Rong-rong Zhu, Da-wei Hu (2012) “A Multi-objective Optimization Model for the Cold Chain Distribution Center Location Problem”, Logistics Technology, 31(01), pp.1-4. Sen Liu,Felix T.S. Chan, S.H. Chung (2011) “A study of distribution center location based on the rough sets and interactive multi-objective fuzzy decision theory”, Robotics and Computer-Integrated Manufacturing, 27(2), pp.426-433. Nick T. Thomopoulos (2016) Elements of Manufacturing, Distribution and Logistics: Quantitative Methods for Planning and Control. Springer International Publishing: Heidelberg New York Dordrecht London. Yu Rong, Jia-feng Chen, Gong Ya,Zhou Bin (2015) “Research on the Location Selection of the Logistics Distribution Center Based on Fuzzy Comprehensive Evaluation Method: A Case Study of Debon logistics 141
Rev. Téc. Ing. Univ. Zulia. Vol. 39, Nº 4, 133 - 142, 2016
in Taixing City, Jiangsu Province”, Scientific and Technological Management of Land and Resources, 32(04),pp.1-8. Ya-meng Wang, Zhang Chen (2015) “Logistics center location decision analysis research”, Journal of Civil, Architectural &Environment Engineering, 37(08), pp.1-5.
142