An evolutionary approach to product line design and pricing
Date of Issue2016-10-14
School of Mechanical and Aerospace Engineering
With the fragmentation of mass markets in many industries, organizations are increasingly competing on product lines as a marketing strategy instead of focusing on single product. By offering product lines, companies can cover a broader range of market segments but yet do not incur significant cost disadvantages because of economies of scale in material and production processes. Product lines usually evolve in response to market or technology changes with new products phased in and poor-performing products phased out. The objective of this research is to develop an evolutionary product line design methodology in adapting product lines for improved profit in order to survive and proliferate in a competitive marketplace. Some companies may develop new products and select elements for new product lines in sequence. Some companies may conduct these two activities simultaneously. In this regard, this thesis studies two versions of integrated decisions of product line adaptation: 1) including new product design, product phase out and pricing, and 2)including product phase in, product phase out and pricing. A mixed logit (ML) discrete choice model and a time-driven activity-based costing (ABC) model are respectively developed to quantify the demand and cost implications of product line adaptation. Quasi random simulated maximum likelihood estimation is developed to estimate customer preferences for the mixed logit model. The decisions regarding changes in product attributes, mix and prices are modeled as a mixed integer or a mixed integer-discrete-continuous non-linear programming problem with the goal to maximize profit. A bi-level optimization procedure, combining genetic algorithm (GA) and differential evolution (DE), is developed for problem-solving. The proposed methodology is illustrated with an example of mobile phone product line design. It is indicated that the adaptation of product mix and prices improves a product line’s profit and that updating product mix and prices simultaneously outperforms updating either one separately. The case example has demonstrated the feasibility of the proposed method for product line adaptation. The proposed methodology can help develop profitable product lines in competitive market. Traditional choice models assume that consumers have well-defined preferences and are not influenced by additional information. However, consumers do not always behave rationally in making a purchase decision. A behavioral choice model is developed to employ reference dependence, diminishing sensitivity and loss aversion in assessing the value of each attribute. The behavioral choice model is illustrated with another survey data of the mobile phone, in which each attribute possesses more levels to be more suitable for parameter estimate. It outperforms logit choice model with a larger log-likelihood value at convergence and a smaller squared error in the case study.