Study On Some Key Technologies Of Multi-Agent Based Software Enterprise Human Resource Management System

Human resource management(HRM) is the key to reach the enterprises'strategic goal, and the Human resource management system(HRMS) is the efficient tool for HRM. This dissertation focuses on HRMS in software enterprises and an intelligent HRMS is established based on Multi-Agents technologies. An extensive study on the system's crucial technologies is carried out as well. The main contents and achievements of the dissertation are as follows:1. Taking the software enterprises as the background, the model and process of HRMS are proposed. As a result, a HRMS architecture based on the Multi-Agents is constructed. The architecture meets the needs of all agents, which act at different levels of the system and have different functions and styles. The universal architecture of the Agents is also described.2. Aiming at the multi-Agent system's demand for domain ontology, the definition, building methodology and research situation of ontology are introduced. On top of the analysis of the current ontology's building methodologies, the HRMS ontology's building methodology is presented. According to the methodology, the HRMS ontology is constructed and used in the skill management.The skill's match is also described and proved by a diagram.3. The human resource scheduling model in software project is built. In order to improve the practicality, a skill proficiency parameter is introduced. Following the process type Multi-Agent Scheduling Strategy (MASS), a Genetic Algorithm that is to solve the human resource scheduling problems is designed which is sealed into the scheduling Agent. The human resource scheduling is fulfilled through the interaction of the Agents. The definite structure of the scheduling Agent is described and the validity of Genetic Algorithm is proved by an example.4. In order to realize the human resource's load balance, a human resource redistribution model based on the marketing mechanism is presented. The model architecture is presented and the employed techniques are explained, with special emphasis on the specific multi-criteria negotiation and Agent coalition formation protocols. The multi-criteria coalition negotiation algorithms and strategies are developed. And an application example is described to illustrate the reached solution for a concrete scenario. 5. The Agents are realized in Grasshopper platform and the ABIHRMS framework is built based on J2EE. The system's runtime environment and example screens are presented. This system has been applied in Jiangyin Longyue information techonology Co., Ltd.