Although the issue of uncertainties in cost model parameters has been recognized as an important aspect of life-cycle cost analysis, it is often ignored or not well treated in cost estimating. A simulation approach employing kernel estimation techniques and their asymptotic properties in the development of the probability distribution functions (PDFs) of cost estimates is proposed. This eliminates the guess work inherent in current simulation based cost estimating procedures, reduces the amount of data sampled and makes it easier to specify the accuracy desired in the estimated distribution. Building energy costs can be reduced considerably if air duct systems are designed for the least life-cycle cost. The IPS-Method, a simple approach to HVAC air duct design is suggested. The Diameter and Enhanced Friction Charts are also developed. These are charts that implicitly incorporate the LCC and are better than the existing Friction Chart for the selection of duct sizes. Through illustrative examples, the ease and effectiveness of these are demonstrated. For more complex designs, a Segregated Genetic Algorithm (SGA) is recommend. A sample problem with variable time-of-day operating conditions and utility rates is used to illustrate its capabilities. The results are compared to those obtained using weighted average flow rates and utility rates to show the life-cycle cost savings possible by using this approach. Although life-cycle cost savings may be only between 0.4% and 8.3% for some simple designs, much larger savings may occur with more complex designs and operating constraints. The SGA is combined with probabilistic cost estimating to optimize HVAC air duct systems with uncertainties in the model parameters. The designs based on the SGA method tended to be less sensitive to typical variations in the component physical parameters and, therefore, are expected to result in lower balancing and operating costs.