本文研究诸如钢铁、有色金属、基本化工等一类行业的空间配置。由于韦勃、廖施的理论的一些前提条件在这里不存在，使得他们的研究方法显得十分不足。本文作者建立了具有对偶的线性规划模式来解决这类行业的空间配置。这一对模式的解给出了具有最小行业总成本的配置方案和与之相应的产品和原料在各个区位的影子价格。尤其是作者依据空间价格均衡找出了原模式与对偶模式的解之间的关系，这些关系是行业空间配置优化的效率条件。

Locational researchers are interested in not only the principles of location choice for an individual firm, but also the spatial allocation of an industry. This paper attends to study the production allocation in space for such industries as iron and steel, nonferrous metal, and organic synthesis etc, in each of which firms are quite big in scale, but the total number of them is small, thereby locational choice for each of them has a quite great effect on the others. Otherwise, the demands of these industries distribute unequally in space. For this kind of industries the analytical methods used by Weber and Losch appear to be invalid because their presumptions do not exist here. Weber’s theory needs to be given the material and product sale sites for each one, therefore it does not consider the spatial competition in the same industry. Losch’s theory ignores the unequality of resource and product demand distribution in space. The author of this paper proposed a linear model with its dual for explaining the spatial allocation of this kind of industries.First, the author assumes that there are many production sites available for an industry which has given materials supply and markets. Also he makes capacity b^h_i and unit cost f^h_i of material h in its site i as well as demand amount q_j; of the product in its sale place j presumptively. Simultaneously there are some exogenous variables such as unit product processing cost c₁; of production site i, transport cost rate t^h_ij; of material h from its site i to production site j,transport cost rate t_ij of the product from production site i to its sale place j. Then the spatial allocation of the industry can be reduced to a linear programming problem as follow: Where rh =technique coefficient related to material h;xij=transport amounts of material h from its site i;to production site j;yi =output of the product in production site i;zij=transport amounts of the product from production site i to sale place j.The model is to minimize the total cost including the production and transport both for material and product under the capacity of every material site and the demand quantites of every market. Its solution is composed of the output from every material and product site, as well as transport amounts from material sites to product processing sites, and then to the markets. Based upon spatial price equilibrium the author sets up the dual model of the linear programming. Its objective function is to maximize the industrial economic income obtained from the production and transport of materials and products: Its constraints are as follows:<1> The difference between the product shadow prices in a market and production site is not in excess of the transport cost rate conneected the two places:pj-ωi≤tij <2> The difference between the product shadow prices and its material cost is not in excess of its production cost per unit at every producing site: <3> The difference between shadow price of a material at production site and the economic rent at its site is not in excess of the sum of its unit cost added transport cost rate connected the two places: where pj= shadow price of the product at sale place j wi=shadow price of the product at production place i vjh=shadow price of material h at product processing site j uih=shadow price or economic rent of material h at its site i From solving the dual model, what can be got are the shadow prices of the materials and product at respective places. Thus, the solutions of the primary and its dual model give out the industrial distribution in space with the minimum-total-cost and shadow prices of the materials and product at respective locations.Afterwards, the author seeks the relationships between solutions of the two models from spatial price equilibrium, and gets following results: <1> The product is transported to a sale place from a production site only when the difference of the shadow prices offsets the transport cost rate between the two places:(pj-ωi-tij)zij = 0<2> The product is produced only when the shadow price blances its per unit cos