The purpose of supply chain management is to plan, coordinate, and control a net- work of facilities that deals with the supply of raw materials, the transformation of these into intermediate and eventually finished products, as well as the distribution of those to the customers (see, e.g., Simchi-Levi et al. 2007). This is a very complex task even for relatively small supply chains. One approach to reducing the complexity is to split the supply chain planning problem into subproblems that resemble the hierarchical structure of the decisions: The higher-level decisions impose constraints on the lower-level decisions, whereas the lower-level decisions provide the necessary feedback to evaluate the higher-level decisions. This is one of the main ideas behind hierarchical planning (see, e.g., Hax and Candea 1984, Bitran and Tirupati 1993).
Early work on hierarchical planning motivated by planning and scheduling prob- lems can be found in Hax and Meal (1975) and Bitran and Tirupati (1993). Since then, hierarchical planning has been applied to a wide range of problems. A recent example discussing a hierarchical approach to supply chain management is given by Miller (2002).
To describe the hierarchical links between the different planning problems, we classify the planning problems according to the framework proposed by Anthony (1965). He distinguishes between decisions on strategic, tactical, and operational levels. The strategic level considers long-term decisions, the tactical level supports medium-term decisions, and the operational level deals with short-term decisions.
The actual length of the time horizons on different levels can vary between different kinds of industry. While a strategic model in the oil and gas industry can have a time horizon of up to 50 years, a strategic model in the semiconductor industry, where the life span of the products is much shorter, can be only 5 to 10 years. Similar differences can be found on the tactical and operational levels.
Figure 4.1 shows the different elements of our planning hierarchy. On the strate- gic and tactical level we plan for the supply chain as a whole. As volume flexibility and storage flexibility are installed primarily at this level, we need the whole supply chain perspective to determine where and how flexibility is needed. On the operational planning level we include a facility planning level in addition to the supply chain.
The main reason for this is that most of the operational decisions will ultimately be implemented at the facility level, so we will exploit flexibility here.
Saunder August 2, 2010 12:44 AU945X˙C004
50 Peter Sch ¨utz et al.
Strategic Supply Chain Planning
Tactical Supply Chain Planning
Operational Supply Chain Planning
Operational Facility Planning
Operational Facility Planning
Operational Facility Planning
Operational Facility Planning Figure 4.1 Supply chain planning hierarchy.
The upper-level planning problems will not only impose constraints on the lower-level problems in terms of capacity or target production volumes, but also determine the type and amount of flexibility available at the lower level. Usually, flexibility is installed at the higher level and exploited at the lower level. The results from the lower-level problems being used to evaluate the decisions made at the higher level.
4.2.1 Strategic Planning
The highest hierarchical level in Anthony’s framework is strategic planning. In short, strategic decisions define the objectives of the organization and plan the resources used to obtain these objectives. Typical examples of decisions on the strategic level are supply chain design and facility location decisions.
Decisions at this level have a significant impact on the supply chain for a relatively long period. Once implemented, it may be impossible, or at least very expensive, to revise strategic decisions over a certain period. This implies that strategic planning has to consider a planning horizon of several years into the future. The degree of uncertainty increases with the length of the planning horizon. Models supporting strategic decisions should give solutions that are flexible when it comes to adjusting decisions according to future events in the market. Often, stochastic models are used that take into account uncertainty in, for example, market demand or market prices.
Strategic models usually aggregate future demand to reduce the number of time periods in the model. This approach captures uncertainty in the long run, e.g.,
Saunder August 2, 2010 12:44 AU945X˙C004
Using Hierarchical Planning to Exploit Supply Chain Flexibility 51
changes in market size or market shares. Short-term variations, however, disappear from the dataset. Neglecting these short-term variations will underestimate the need for capacity.
Sch¨utz et al. (2009), for example, examine the case of redesigning a supply chain given uncertain demand. They show that a stochastic model will install more capacity than the deterministic expected value problem where uncertain demand is replaced by its expected value. Also, modeling short-term demand variations instead of using aggregated demand levels results in a higher demand for flexibility. The stochastic model is able to value and install volume flexibility (i.e., excess capacity).
This flexibility can be exploited in the future, as the production volumes can be increased depending on realized demand.
4.2.2 Tactical Planning
The next level in the planning hierarchy is tactical planning. According to An- thony (1965), the purpose of tactical planning is to ensure the efficient and effective utilization of the organization’s resources. The planning horizon is usually several months long and should cover a seasonal cycle (Hax and Candea 1984, Miller 2002).
The tasks include balancing supply and demand, assigning production volumes to different plants, planning capacity utilization, workforce planning, and inventory planning, among others.
Traditionally, tactical planning has focused on how to adjust the production system in order to meet fluctuating customer demand (Hax and Candea 1984, Swaminathan and Tayur 2003). These adjustments can be made, for example, by changing the workforce, or building and depleting inventory. An early example of combining inventory control and production smoothing in aggregate production planning can be found in Winters (1962). Inventory is often carried to hedge against variations in demand. These variations in demand can be known in advance (e.g., seasonal patterns) but can also be uncertain.
Aggregate production planning exploits volume flexibility in the supply chain by allocating production volumes and adjusting the workforce at the different produc- tion facilities. Using storage flexibility by building and depleting inventories allows not only for smoothing capacity utilization, but also to prepare for seasonal variations in demand. Focusing only on the production system, however, ignores the possibility of balancing supply and demand by actively managing customer demand. Combin- ing aggregate production planning with methods from revenue management (see, e.g., Talluri and van Ryzin 2004) introduces additional flexibility: As an alterna- tive to building inventory or reducing the rate of production, demand for finished products could be increased using a time-limited marketing campaign. This type of flexibility is very important for supply chains dealing with perishable products or other products with a limited shelf life.
Saunder August 2, 2010 12:44 AU945X˙C004
52 Peter Sch ¨utz et al.
4.2.3 Operational Planning
On the operational planning level, we focus on planning activities for the next few weeks. Capacity limits and often also supply and demand are fixed for at least the beginning of the planning horizon. The problem is to determine the best production plan for the near future. The level of detail must be very high on this level to be able to create a plan that takes into account all real-life constraints of the production processes. As a consequence, it is often impossible to solve such a problem for the whole supply chain in a reasonable period of time. We therefore suggest using moderately aggregated data for operational supply chain planning and detail these plans further by means of operational planning for each facility.
Flexibility at the operational level is provided mainly by the possibility to allocate specified production volumes to different facilities or to use different bills of material in the production process. In addition to this operational flexibility, we can exploit the flexibility designed into the production network at the higher levels of the planning hierarchy. Considering short-term variations in, for example, demand during the supply chain design phase increases the amount of volume flexibility available in the supply chain. We can therefore react to large demand variations by depleting seasonal inventories of raw materials and increasing the production volumes at different production facilities in order to satisfy observed demand.
Sch¨utz and Tomasgard (2008) studied the impact of different types of flexibility on operational supply chain planning under uncertainty. Their results show that operational flexibility becomes more important in case little or no volume and storage flexibility are installed. They also showed that the planning process is not affected by uncertainty if enough flexibility is available in the system.