An urban land-use plan-design model is cast as a multiple-criterion optimization problem and is solved by using a generalized interactive technique. Three different objectives are selected for use in a representative formulation: (1) the minimization of land-development costs, (2) the maximization of residential accessibility, and (3) the minimization of energy consumption in transportation. It is found that interactive multiple-objective programming represents an efficient way of generating compromise land-use plans which provide satisfactory levels for the various objectives. The approach is illustrated for a hypothetical planning problem for Germantown, Wisconsin, USA. IntroductionDespite considerable research on urban-simulation models, such as the National Bureau of Economic Research (NBER) model and the descendants of the Lowry model (for example, Goldner, 1968;1972;Crecine, 1964;Batty, 1976), there are few operational versions available and there seems to be only faint hope that a satisfactory model will appear within the next few years. Existing models are so complex, time consuming, and expensive to operationalize that their use in practical land-use planning situations is difficult. Land-use plan-design models offer an alternative device to aid planners in developing land-use plans for urban areas, especially for initial sketch planning. These models treat urban land-use planning as a design problem; they allocate future land-use requirements to the zones of an urban area so that some specific criterion function is optimized.Various formulations of land-use plan-design models include linear-allocation models (see Ben-Shahar et al., 1969;Brotchie et al., 1973;Ochs, 1969;Schlager, 1965;Stewart and Greco, 1970), quadratic-allocation models (see Brotchie et al., 1973;Gordon and MacReynolds, 1974;Lundquist, 1974), and hybrid models such as TOPAZ (see Dickey and Najafi, 1973;Dickey and Sharpe, 1974;Sharpe et al., 1973). All of these authors specify mathematical programs to allocate future land-use growth so that a single objective, usually development costs, is minimized. Land-development costs and transportation objectives are included in a few models but they are reduced to monetary costs in order to specify a single, composite, objective function. Dickey and Sharpe (1974) explicitly recognize the existence of multiple objectives, but land-use plans that optimize each objective are independently determined. No attempt is made to determine a compromise land-use plan even though the various objectives are clearly in conflict.By its very nature land-use planning is concerned with the resolution of multiple, conflicting, often incommensurate and occasionally even intangible, objectives. Decisions about land-use plans must be weighed against the social, economic, political, environmental, and aesthetic needs of urban residents. The criteria include accessibility to opportunities, land-development costs, various transportation-related factors such as air quality, energy consumption, and average travel times, and ...

In this paper a recursive version of a dynamic road investment problem is developed by extending the Ellet-Walters developmental model of transportation so as to account explicitly for the effects of road investment on the supply and demand of the port hinterland's product. An explicit lagged supply function that depends upon commodity price and road penetration is related to a given linear demand function, assuming annual market clearance. A recursive formulation of a cobweb-like dynamic market is thus obtained. Secondly, we derive the optimal roadpenetration policy of a road-planning authority faced with a given cost function and with this form of market dynamics. The road-building policy is constrained both by long-run goals and by short-run market fluctuations so that there are no short-run disbenefits from development. We show that these constraints hold for the general case of downward-sloping demand functions and price and of road-sensitive supplies. The results are then extended to cases where farmers have differing views of future transport availability, indicating the importance of people's perception of development. The paper ends by considering the conditions under which development does not take place. The results imply that an appropriate range of technological change occurs, and that it is related to the existing level of development and the inherent stability of the market.

In this paper the determinants of internal migration in Kenya are analyzed on the basis of a human capital model. Explanatory variables included in the specification are both economic (wage rates and employment rates) and noneconomic (for example, population density and educational attainment). Also incorporated are variables which reflect intervening opportunities. These variables are defined as distance-weighted averages of the variables in all of the districts in Kenya except the origin and destination districts. The econometric results show that destination variables are important determinants of internal migration, as is distance between the districts. Further, the variables for the intervening opportunities add significantly to the explanatory power of the model.

Regional transport investment planning has traditionally assumed that virtually all the relevant information necessary to successful planning is known with complete certainty. More realistically, these investment decisions can be viewed as taking place in an environment where most of the necessary information is not known with any degree of certainty and where unpredictable events might alter the suitability of investment decisions made at a previous point in time. In such a decision-making situation, it is reasonable to select transport investment projects which have the best chance of being good investment choices in many different future operating environments. Though this may result in a less than optimal investment package in the future which is actually realized, it is most unlikely that the chosen investment package will be a poor choice. Strategies for efficient decision making in an environment of uncertainty can be derived by simple modification of models used in studies which assume a complete certainty of outcomes. Several such strategies are illustrated for a road investment planning problem in South Sulawesi, Indonesia.NVESTMENT in transport network infra-

The transportation requirements of an urban area are, in part, a function of its land use pattern. Reducing travel in an urban area by effective land use design appears to be a worthwhile public policy objective. The planning model derived in this article uses as its core the matrix version of the Lowry model developed by Garin. The model allocates basic employment to the zones of an urban area in order to minimize work and service trips. A hypothetical planning problem for Metropolitan Toronto illustrates use of the model.