Supplier Selection Group Decision-making Modelling And Multi-source And Multi-period Procurement Optimization

Supplier selection in supply chains plays a more and more important strategic role in reducing cost, increasing efficiency and improving service. In view of resource constraints and uncertainties of demand and supply, large enterprises and project construction units usually select several suppliers according to multiple objectives or attributes at first, and then allocate orders among them. In order to reduce the impact of supply chain uncertainties and promote cooperation relationship, suppliers and demanders usually sign some bilateral contracts with each other. As an effective method for disposing the uncertainties in supply chains, setting up safety stock plays an important role in reducing cost and improving customer service. So developing some methods supporting supplier selection, order allocation and safety stock placement is of important academic and practical significance to enrich and perfect supply chain management and decision analysis theories and promote effective operation of supply chains.This paper summarized and analyzed the situation and development trend of researches on supplier selection, order allocation and safety stock placement, and then did some research on supplier selection group decision-making modeling, and multi-period and multi-supplier order allocation and safety stock placement in the presence of stochastic demands, resource constraints and supply chain contracts.Firstly, several new multi-attribute group decision-making models were developed for supplier selection. The method based on group's ideal solutions which extended the technique for order preference by similarity to ideal solution (TOPSIS) to group decision-making, substituted group's ideal solution for group utility and was beneficial to differentiate the alternatives. The TOPSIS with experts' synthetic weights under ordinal preferences in which a new expert's synthetic weights setting method was proposed and this methodology can avoid the conundrum of set the weighting coefficient when synthesizing subjective expert' weights with objective ones to synthetic weights, generalizes TOPSIS to ordinal preferences and eliminates rank reversal. The method minimizing the distances of ordinal preferences in which two distance functions for ranking vectors were defined and one of which was been proved to satisfy the conditions proposed by Cook and Seiford in social selection theory and the methodology extends Cook-Seiford social selection function to multi-attribute decision-making with weights and may avoid non-uniqueness of the solutions for the same decision-making problem.Secondly, multi-supplier and multi-period order allocation optimization under some supply chain contracts was investigated. The order allocation problems in multi-supplier and one-buyer scenario where every supplier specified the minimum and maximum of the quantity purchased each period and the total minimum purchased over the predetermined plan horizon were researched as follows. Under the conditions of certain demands and no short a model minimizing the sum of purchase cost, inventory cost and transportation cost of the buyer over the plan horizon was developed and for which a solution integrating multi-dimensional dynamic programming and heuristic algorithm was proposed. For the scenario where demands in every period are stochastic and independent and the stockout is permitted, an optimization model was developed and which minimized the expected value of the total cost including purchase, inventory, stockout cost and salvage value of the ending inventory for the buyer over the plan horizon. At first, the optimal procurement policy in single period was obtained according to the newsboy model after the order constraints and multiple suppliers were considered. Then the heuristic policy for multi-period procurement was derived from the one in single period. Some simulations were made to show the effect of the parameters on the optimal policy and the minimum cost. This work extends the total minimum contract to the stochastic and non-stationary demands and multiple suppliers setting, and extends newsboy model to the setting with multi-supplier, multi-period, resource constraints and supply chain contracts.Finally, multi-period safety stock placement with stochastic demands and resource constraints was dealt with. For a one-supplier and one-buyer supply chain with stochastic demands, order lot size constraints from the supplier and the buyer's storage capacity constrains, an optimization model was developed and which minimized the expected value of the sum of order, purchase, inventory, and stockout cost in unit time between two successive order when the buyer used (R, Q) policy. A solution based on particle swarm optimization (PSO) was proposed and some case study and simulations were made. The results showed that the optimal reorder point and the optimal order quantity is dependent on the product price and that holding no safety stock may be optimal in some cases. Then the incremental quantity discounts was taken into account and a PSO-based algorithm was developed to obtain the optimal (R, Q) policy and selection supplier for the buyer. The works correct some mistakes in published literatures on (R, Q) optimization and extend safety stock research to the settings with resource constraints and supply chain contracts. These results can provide a theoretic basis for the buyer to select the suppliers with different ordering cost, product price, order quantity constraints and quantity discount menu.