Abstract: The multi-attribute decision making problems that the attribute weights（interval numbers） are partly known and the attribute values are numeric are investigated.By building single-goal linear programming models，locally optimal weight vectors of alternatives are respectively obtained.According to minimize deviation of globally optimal weight vector and locally optimal weight vector，goal-programming model are established，then by solving the single goal-programming that transferred goal-programming model obtained above into，globally optimal weight vector is obtained in this paper.And a new method based on two-step goal-programming model for multi-attribute decision-making is proposed.Finally，a practical example is given to illustrate that the method is feasible and effective.

Key words: multi-attribute decision-making, goal programming, weight vector, ranking priority

摘要： 研究了只有部分权重信息（区间数），属性值为实数的多属性决策问题。通过建立单目标线性规划模型求出各方案的局部最优权重向量，通过将各方案的局部最优权重向量与全局最佳属性权重的偏差最小化，建立多目标优化模型，并将其转化成一个单目标优化模型计算出全局最优属性权重向量，从而得到了一个两阶段的多属性决策方法。通过实例说明了该方法的可行性和有效性。

关键词: 多属性决策, 目标规划, 权重向量, 排序