Study On Optimization Problems Of Scheduling And Inventory In Production Logistics

Compare to circulation enterprise's logistics, manufacturer's logistics ususally involves more complex envirenment and faces more constraints, and it is more difficult and more important in logistics research. This Doctoral Dissertation addresses the topic of inventory control and scheduling of manufacturer's logistics.(1) single item's production and inventory problem when purchasing is consideredDuring studying on single item's production and inventory problem when purchasing is considered, such problems are studied mainly:(a) optimization model of production lot when producer buys material in bulk; (b) optimization model of production lot when producer buys material in bulk and item's stockout is allowed; (c) optimization model of production lot when material has discount and item's stockout is allowed; (d) optimization model of production lot when producer buys material in bulk and item is perishable; (e) optimization model of production lot when material has discount and item is perishable.The supplier and the demander are equal in most of the time. The demander is not dominated, and it will adjust its inventory and purchasing policy according to the supplier's supply policy. Moreover, even if the supplier adopts unique price policy, the demander will select a suitable purchase quantity according to its own characteristics of demand, since usually there is a long way to go between the supplier and the demander, and logistic cost is closely interrelated with purchase quantity, and the purchase quantity may influence production lot. So, these models may provide some reference and guidance to relevant enterprises'realistic questions such as inventory and purchasing. (2) multi-item's production and inventory problemThis Dissertation offers new algorithms to solve multi-item's production and inventory problem which is also called economic lot scheduling problem (ELSP). The dissertation, by using Pareto principle to analyze ELSP, offers the difficult point and the important point of solving ELSP, so the new algorithms can obtain high quality solutions.1) Current algorithms are analyzed and commented.2) When power of two (PoT) strategy is adopted (item's production frequency or multiplier value is restrict to power of 2), basing on pin-packing thought, this dissertation offers optimal solution after simple computing by finding out difficult items and changing pin-packing object to "as possible as close to but not beyond B" (extended basic period approach) "at most to the given height" (time varying approach) and obtaining production multiplier or frequency vector with the minimal cost.3) When non power of two (nPoT) strategy is adopted (item's production frequency or multiplier value is not restrict to power of 2), by finding out important items and difficult items, this dissertation offers the algorithms'improving thinking:during obtaining the high quality multiplier vector or production frequency vector, more attention should be paid to such items whose daily cost is high; during loading items, more attention should be paid to difficult items.By analyzing, this dissertation extends the necessary condition when extended basic period approach is used to solve ELSP, and makes the model more close to real world.By multiplying extended coefficiency (EC), items'production multiplier value (or production frequency value) may be more harmonious, and scheduling is easy. By obtaining high quality production multiplier vector and basing on extended basic period approach, an near optimal algorithm is offered, and it is fit for such conditions when utilization level is low.A near optimal algorithm is offered when utilization level is high. By rounding off item's production frequency to PoT or the media value of 2 adjacent PoTs, item's production frequency will be harmonious, and a high quality scheduling can be quickly obtained.4) A near optimal algorithm of ELSP is offered when shelf life is considered. When utiliaztion level is high, usually we can only use time-varying lot size approaches to solve it. Since it is difficult for production frequencies to be harmonized by multiplying EC, so item's production frequency will be rounded to near even values. When utilization level is low, we can use period approach to solve it.New algorithms offer the best solutions to date in benchmark examples.