Planning disassembly with stochastic yields

Abstract

In many production settings, chemical and electronic component manufacture among the most well known, yield uncertainty complicates the planning by marring the input-output relationship with uncertainty (Yano and Lee (1995)). This supply side uncertainty compounds the more typically thought of demand uncertainty. Product recovery environments exhibit even more uncertainty than traditional manufacturing particularly on the supply side as a result of the uncertainty surrounding both the amount as well as the quality of the cores (see Guide (2000), Guide et al. (2000), and more recently Toffel (2004)). This latter aspect, quality uncertainty, results in the fact that when we disassemble a certain core, we are unsure how many good quality leaves we will obtain for remanufacturing. With respect to demand uncertainty, one can mention that it is often the case that the responsibility for product availability belongs to one division (e.g. distribution) whereas remanufacturing operations are the responsibility of another division. When this is the case, the latter receives a fixed order from the former and must meet this specified demand, a so-called remanufacture to order system (see Guide et al. (2003)). Incorporating yield uncertainty into planning, as mentioned in Yano and Lee (1995) while difficult, allows for (1) more effective operational control over a system, (2) improved tactical decision making, such as supplier selection and process improvements, as well as (3) better utilization of capacity in the medium to long term. In our setting, we are mostly interested in more operational matters, lower average costs and more robust planning.