UC Berkeley

This dissertation contributes to the sustainability literature by incorporating a systematic analysis to provide guidelines to sustainable design. During the design phase, there is a great potential to manage the environmental impacts of the whole product life cycle. However, decision making in the design phase is usually complicated. Therefore, a holistic system analysis with the consideration of relevant factors in the whole product life cycle is needed. In this dissertation, a life-cycle thinking framework is applied to solve the supply chain network design problem and the product durability design problem.

In Chapter 2, an integrated supply chain network design problem is studied. We built optimization models to minimize the carbon emission and/or the cost of a three-tiers supply chain network. The carbon emissions associated with major supply chain activities, including transportation, warehouse construction, and warehouse operations, are characterized and included into the optimization models. The decisions in the models include transportation mode, shipping frequency, warehouse locations, warehouse capacity, and warehouse-retail store assignment. We transformed the problems into second-order conic programs to reduce the computation time. Several model variations, including multi-objective program and bi-level optimization model, are also studied to understand the trade-offs between cost and carbon emissions and provide sustainable guideline for supply chain design in real-world scenarios.

In Chapter 3, the product durability design problem is analyzed. We constructed a mathematical model to understand whether Design for Durability strategy is always good for the environment or not. In the study, the energy consumption during the manufacturing phase, the use phase, and the remanufacturing phase is considered. The bass diffusion model is adapted and modified in the study to enable the construction of the relationship between new manufactured products and refurbished products. Analytical results confirm the common beliefs that a more durable product consumes less energy in the special case with no repeat purchases. A numerical analysis for the general setting identifies the main factors of the optimal product life when we want to minimize the average energy per item per unit time. We also provide exploitation on the various trade-offs of increasing durability to better understand optimal product durability design.

In Chapter 4, we discussed a promising future direction on incorporating Life-Cycle Assessment (LCA) uncertainty into decision making processes. LCA uncertainty has gathered great importance in recent years. Researchers have argued that, without incorporate the uncertainty analysis, the decision made based on a single and deterministic LCA result may be unreliable. Quantitative uncertainty analysis is needed to incorporate the uncertainty level into decision making processes. However, due to the time and cost constraints, it is not always feasible to conduct a complete LCA study with uncertainty analysis. It becomes a challenge to quantify the uncertainty level of a LCA study when only limited secondary data is available. We proposed a future work direction that could remedy the drawbacks of current methodologies.