摘要

目前，中国正处于大力发展城镇化的改革进程中，2015年全国有40余座城市进行修建城市轨道交通，规划线路长度是2400公里，投资规模近7000亿。地铁车站招投标作为建筑工程的初始阶段，对工程建设具有深远影响，目前我国绝大多数城市地铁车站招投标依旧陷于低价中标的困境之中，招投标工作的特点之一是实践性强，需要依赖专家经验和知识，而建筑业人才的高流动性使得这些知识，尤其是其中的隐性知识的积累缺乏可持续性，如何将过去和现在发生的地铁车站招投标案例有效的收集、保存、学习、重用是摆在我们面前及待解决的问题。

近年来知识管理在工程管理领域越来越受到专家和学者的重视，其中基于案例的推理是人工智能领域的重要研究方向，该方法解决问题的基本思路与招投标工程造价估算实际工作中工程信息化、智能化的思想不谋而合。

鉴于此，本文提出了一种基于知识管理的招投标阶段地铁车站工程造价估算的方法，分析了招投标工程造价问题和基于案例推理二者结合的必要性和可行性，综合运用向量空间、文本挖掘、相似聚类等方法建立基于案例推理的招投标工程案例知识库系统，并进行了实证分析，实现了基于案例推理的招投标工程造价的估算，通过工程造价估算结果与实际发生造价相比，偏差范围约在-7%—8%之间，具有一定参考价值。

关键词：知识管理案件推理向量空间文本聚类造价估算

Abstract

At present, thereform of urbanization is developed in china. There are more than 40 citiesconstructing urban rail transit during 2015 including 2400 km planning linelength and the scale of investment of nearly 700 billion. Metro station has aprofound impact on the construction as the initial stage of the biddingconstruction projects. Currently the vast majority of urban subway stationsstill troubled in low bid. Bidding work has strong practical feature and needsto rely on the expertise and knowledge. However, these knowledge especially theaccumulation of tacit knowledge lack sustainability because of the highmobility of construction talent. So how to efficiently collect, study and reusebidding cases in metro station occurred in the past and present is the problemsto be solved.

In recent years,the knowledge management is more and more attention by experts and scholars thefield of project management, of which case-based reasoning is an importantresearch field of artificial intelligence. The basic ideas of the solution tothe problem have the same view happen to the actual engineering work,information, intelligence ideas in cost estimate in bidding project.

In view of this,we propose the cost estimate method during bidding in subway station based onknowledge management issues existing in bidding project Based on the casereasoning of case knowledge bidding system, the paper integrated applicatevector space, text mining, clustering similar knowledge of algorithms andconstruct the algorithms using the above theories and methods of case analysis.Then the paper realize a preliminary study based on case-based reasoningbidding project cost estimation problems, Compared with the actual cost of theproject, the deviation range is about -7%--8%, which has a certain referencevalue.

Key words:knowledge management, reasoning based case, vector space, text clustering