Inventory is the blocked working capital held in the form of raw material,
semi-finished products, spare parts, finished products, etc. at different stages of the Supply Chain
(SC) for future use or sale. SC managers in the manufacturing environment constantly
face challenges from inventory planning to determine the appropriate inventory level,
service level, etc. at each stage of the SC in order to minimize the total relevant system-wide
cost in today's rapid and violent market place. Rangaraj
et al. (2009) mention that "organizations need to continue their efforts to drive down inventories as far as
possible simply because the presence of inventories is a symptom or manifestation of long
lead times and long lead times only serve to increase the bullwhip effect" (p. 51). Monczka
et al. (2004) mentioned that "the wrong reasons for carrying inventory are poor
quality and market yield, unreliable supplier delivery, extended order cycle times, inaccurate
or uncertain demand forecasts, specifying custom items for standard applications,
extended material pipe lines and inefficient manufacturing processes" (p. 402). The
inventory investment may add value by reducing costs in areas like logistics, manufacturing, etc.
in the manufacturing environment. Therefore, a trade-off is essential between
cost implications and potential benefits of maintaining inventory in the SC when
making inventory decisions.
From the time Harris (1913) developed Economic Order Quantity (EOQ)
concept, researchers, academicians and practitioners have been continuously focusing on
inventory planning in various environments under different operating parameters and
modeling assumptions. The EOQ, Economic Production Quantity (EPQ) and Economic Order
Interval (EOI) methods are used to determine lot sizes for continuous and independent demand
items, considering that demand occurs with certainty at a constant rate. Various approaches are
also developed to handle varying demand rates. This is the case for components and
subassembly (i.e., dependent demand products) in Material Requirement Planning (MRP) systems
or finished products in a Distribution Requirement Planning (DRP) system. The
simplest method to handle varying discrete demand for lot sizing is lot for lot ordering, where the
order for each period is the exact quantity in that period and it is rarely used in practice. Some
of the other approaches are period order quantity methods, dynamic programming to
determine the optimum varying order size by Wagner and Whitin (1958), Silver and Meal
heuristic algorithm (Silver and Meal, 1969; and Silver and Meal, 1973) and marginal cost
algorithm (Groff, 1979; and Kicks and Donaldson, 1980). Various researchers use different tools
for inventory planning in different environments. The tools used, but not limited to are
simple calculus, linear programming, nonlinear programming, multi-objective optimization,
heuristics (genetic algorithm, swarm optimization, differential evolutionary algorithm,
simulated algorithm), system dynamics, Just-in-Time, fuzzy optimization and simulation
optimization. Many researchers considered demand to be stationary (Wagner and Whitin, 1958; Florian
and Klein, 1971; Monahan, 1984; Cachon and Zipkin, 1999; and Graves and Willems, 2000)
for inventory planning. Lau and Lau (1995) studied multi-product multi-constraint
newsboy problem for independent demands. One can find literatures (Clark and Scarf,
1960; Federgruen and Zipkin, 1984; Schmidt and Nahmias, 1985; Cohen and Lee, 1988;
Rosling, 1989; McGavin et al. 1993; Nahmias and Smith, 1994; Thomas and Griffin, 1996; Lee
and Whang, 1999; Timpe and Kallrath, 2000; Jayaraman and Pirkul, 2001; Lee and Kim,
2002; Routroy and Kodali, 2005; Routroy and Sanisetty, 2007; and Routroy and Maddala,
2009) regarding multi-echelon inventory planning. |
