J. Paul Riley

Document Type


Publication Date

January 1970


Introduction: The problems of managing water-resource systems are basically those of decision making based upon a consideration of the physical, economic, and sociological processes involved. These processes are strongly interrelated and constitute a dynamic and continuous system. Any combination of these interrelated and numerous system variables yields a management solution. At Utah State University the problem of investigating system response to various possible management alternatives is being approached by hybrid computer simulation. The concept of simulation is fundamentally simple. Basically, it is a technique of analysis whereby a model is developed for investigating the behavior or performance of a dynamic prototype system subject to particular constraints and input functions. The model behaves like the prototype system with regard to certain selected variables and can be used to predict probable resopnses when some of the system parameters or input functions are altered. The model represents the interrelated processes of the system by arithmetic and algebraic functions, and by non-mathematical logic processes. Simulation is a useful tool for the creative manipulation of highly complex systems and thus can greatly facilitate appraisals of proposed changes within the corresponding prototypes. In a computer model the various functions and operations of the different parts of the system are interrelated by the concepts of continuity of mass and momentum. These concepts are applied over the particular increments of time and space adopted for the model. It should, therefore, be emphasized that the adquacy of a simulation model is dependent upon the theory and the field data upon which the model is based. Consequently, both the mathematical relationships and the physical input data constitute major contraints in a simulation analysis. In addition, simulation alone does not readily provide optimal solutions. However, each computer run for a set of model parameters and inputs yields and estimate of the probable response of the prototype under the particular conditions established. Thus, through numerous and repetitive computer runs, it is possible to investiage many combinations of system variables and thereby to evolve optimal or near-optimal system design and operation procedures.