The Order And Distribution Scheduling Of General Medical Resource Based On Time-space Network And Chance Constrained Programming

Currently Chinese medical institutions use electronic purchase management systems on the order work of medical resources. However, the important parameters (e.g., the order/transit frequency, the order quantity, and the safety stock) are manually determined by staff based on their experience. By lack of a systematic optimization analysis, This approach lacks a systematic optimization analysis, and rather depends on the staffs subjective judgments. As a result, suboptimal or poor decisions have usually been made. Additionally, in actual operations, the demands of medical resources often change because of disturbances by uncertain factors or incidents, which would possibly make the advance demand forecast inaccurate and cause the original schedule to lose its optimality, or make it difficult to effectively cope with these disturbances to respond the changes under the condition of the existing resources. Consequently, the overall effect of medical operation system will be decreased and the operating cost be increased. Especially for the recent years, since the new healthcare policies such as abolishing drugs addition come into effect, pharmacy, one of the key sources of profit of traditional hospitals, has become the financial burden to hospitals at present Therefore, it is urgent to re-schedufe the order and distribution work of medical resources in order to reduce the operation cost of hospitals.Based on the above analysis, from a hospital’s perspective, this thesis applies mathematical programming methods to construct the medical resources order and distribution scheduling models. The specific research work includes:(1) This thesis systematically considers the demand of resources for every time slot in all hospital departments, the stock capacity and other constraints, and uses the time-space network technique to construct a deterministic single-variety of medical resources order and distribution scheduling model. This model is formulated as a mixed 0-1 integer programming problem with NP-Hard complexity. A simulation-based heuristic algorithm is developed to solve this model coupled with mathematical programming softwares MATLAB and CPLEX.(2) Further, considering the stochastic demand of medical resources and the supplier’s limited supply capacity that occur in actual operations, based on (1), we modify the fixed demand parameter in the deterministic scheduling model as a random variable and change the supplier’s infinite supply level into a fixed one of every week, and use the stochastic chance constrained programming method to develop a stochastic single-variety of medical resources order and distribution scheduling model. We make use of the classical genetic algorithm and program in the environment of MATLAB to solve this chance constrained programming model.(3) Considering the diversity characteristic of demand of resources, this thesis further expands the work in (2) to construct a multi-layer time-space network structure, where each layer of time-space network represents a variety of medical resources, and build a stochastic multi-variety of medical resources order and distribution scheduling model. Genetic algorithm is applied to solve it.Based on the numerical tests and sensitivity analyses, we can conclude that the medical resources order and distribution scheduling models we construct in this thesis could effectively consider the stochastic characteristic of demand of medical resources in actual operations, and obtain the optimal scheduling results of medical resources order and distribution work.