Over the last decade, many previous researches have been conducted on inventory models and control policies. Recently, we have new types of logistics systems for collecting and dispatching materials, entities, energy, and etc. Such CD (Collecting and Dispatching) systems have been discovered in a variety of fields: carbon capture and storage, shale gas production, energy harvesting for wind and solar power, and product recycling. Collected entities arrive along time and are accumulated in storages, released or shipped out by a control policy. They are often dispatched in batch to reduce associated cost and maximize revenue. However, as soon as a dispatch request is made, it could be difficult to become a immediate dispatch by delays due to setup or vehicle travel. The storage itself may undergo overflow or underflow with some incurred costs. The stored entities also induce some holding costs in a storage and operating costs. Therefore, we need to determine when a dispatch should be requested and how much entities should be dispatched. A CD model can be viewed as a reversed model of a conventional inventory control model. In chapter 2, we define and examine the CD systems as the reversed inventory models for collecting and dispatching entities and suggest optimal control policies. We first give some examples of their operations such as the carbon capture and storage (CCS) system, the shale gas production, and recycling to explain the CD systems. Next, we classify the related works into conventional inventory models, dam/reservoir process and reverse logistics, and then reveal their similarities and differences. To compare with reversed inventory models and conventional inventory ones , we propose EDQ/ETQ deterministic reversed inventory model and identify analogies, differences, and relevance with EOQ/EPQ inventory model. We then explain operation of reversed inventory models and control policies. Finally, we extend the dispatch policies for the reversed inventory with overflow allowed. We also compare the characteristics of each reversed inventory model with those of its conventional one’s counterpart. In chapter 3 and 4, we propose stochastic reversed inventory models with (d, Q) and (d, S) policies comparing with (r, Q) and (s, S) policies of the inventory model as a new type of logistics systems for the CD systems, and subsequently determine the dispatch point and the dispatching quantities. In chapter 5, we introduce a new type of newsvendor problem regarding accumulating, dispatching or releasing and shipping out entities in storage by a control policy. We call it reversed newsvendor problem. Collected entities are dispatched by the contracted capacity and it occurs underflow and overflow in the course of dispatch. We define and examine the single period CD systems as reversed cases of the newsvendor model, and also discuss similarities and differences. Controlling issues of the excess amount of inflow happen during the overflow period. And thus, we suggest the overflow control policies; supplementary dispatch and carryover. According to the study, we find a new type of logistics systems and evaluate the similarities and differences between the reversed inventory model and conventional one.
이 논문에서는 수집 및 전송을 위한 새로운 타입의 로지스틱스 문제를 다룬다. 우리는 이 시스템을 수집 및 전송 (CD) 시스템으로 부른다. 이 시스템은 이산화탄소 포집 및 저장 시스템, 셰일가스 생산, 풍력 또는 태양광 발전을 위한 에너지 하베스팅과 재활용 문제 등에 적용될 수 있다. 수집된 엔티티는 비용을 줄이거나 또는 이익을 최대화하기 위해 배치로 전송이 되는데, 전송이 요청이 되고, 실제 전송이 이루어지는데까지 지연시간이 발생한다. 이러한 과정에서, 스토리지는 오버플로우 혹은 언더플로우를 겪게 된다. 스토리지 내에 들어온 엔티티를 보관하기 위해서는 유지비용이 발생하고, 일정 시간이 지난 후에 보관된 엔티티를 전송하게 된다. 이 때, 언제 얼마나 많은 양의 엔티티들을 전송하는지를 결정하는 것이 중요하다. 이러한 수집 및 전송 시스템은, 발생하는 수요를 충족시키기 위한 최적의 주문량을 결정하는 재고모형의 역모델로 볼 수 있다. 우리는 이를 역재고모델이라고 부르는데, 본 연구에서는 확정적 역재고 모형과 기존의 재고 모형과의 유사성과 차이점을 밝히고, 확률적 재고모형으로 확장하여 (r, Q) 모델과 (s, S) 모델과의 비교를 통해 확률적 역재고 모형인 (d, Q)모델과 (d, S) 모델을 제안한다. 또한, 역재고 모형에서 수집된 엔티티에 대해 단일 기간 내에 전송계약을 맺을 경우, 우리는 이를 역 뉴스벤더 모델이라 부르는데, 기존의 전형적인 뉴스벤더 모델과의 유사성과 차이점을 밝힌다. 이 때, 발생하는 오버플로우를 컨트롤하는 전략은 다양하게 존재 한다. 우리는 오버플로우 컨트롤 전략으로 추가 전송하는 경우와 캐리오버 하는 전략을 제안한다.