DARPA ULTRALOG Final Report - Industrial and Manufacturing ...
DARPA ULTRALOG Final Report - Industrial and Manufacturing ...
DARPA ULTRALOG Final Report - Industrial and Manufacturing ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
D<br />
k<br />
: Outflow from high priority queue in the kth interval<br />
λ : Inflow to low priority queue from the outside in the kth interval<br />
k<br />
The system is characterized by the following parameters:<br />
µ<br />
a<br />
: Rate per unit of the schedule cycle at which the low priority queue can be served<br />
µ : Rate per unit of the schedule cycle at which the high priority queue can be served<br />
b<br />
l: The feedback interval in units of the schedule cycle<br />
The following four equations then completely describe the evolution of the system:<br />
A<br />
C<br />
B<br />
k + 1<br />
= Ak<br />
+ λk<br />
− Ck<br />
(1)<br />
Dk<br />
= min( Ak<br />
+ λk<br />
, µ<br />
a<br />
(1 − ))<br />
(2)<br />
µ<br />
k<br />
b<br />
k +1<br />
= Bk<br />
+ Ck<br />
−l<br />
− Dk<br />
(3)<br />
D = min( B + C , µ )<br />
(4)<br />
k<br />
k<br />
k −l<br />
b<br />
Equations (1) <strong>and</strong> (3) are merely conservation rules, while equations (2) <strong>and</strong> (4) model the<br />
constraints on the outflows <strong>and</strong> the interaction between the queues. This model while<br />
conceptually simple, exhibits surprisingly complex behaviors.<br />
Dynamical Behavior<br />
The analytic approach to solve for the flow model under constant arrivals (i.e. λ<br />
k<br />
= λ for all k)<br />
shows several classes of solutions. The system is found to batch its workload even for perfectly<br />
smooth arrival patterns. Following are the characteristics of behavior of the system:<br />
1) Above a threshold arrival rate ( λ ≥ / 2 ), a momentary overload can send the system<br />
µ b<br />
into a number of stable modes of oscillations.<br />
2) Each mode of oscillations is characterized by distinct average queuing delays.<br />
3) The extreme sensitivity to parameters, <strong>and</strong> the existence of chaos, implies the system at a<br />
given time may be any one of a number of distinct steady-state modes.<br />
The batching of the workload can cause significant queuing delays even at moderate occupancies.<br />
Also such oscillatory behavior significantly lowers the real-time capacity of the system. For<br />
details of application of this model in supply chain context, refer to (Kumara et al. 2003).<br />
5.1.2 Managerial Systems<br />
Decision-making is another typical characteristic in which the entities in a supply chain are<br />
continuously engaged in. Entities make decisions to optimize their self-interests, often based on<br />
local, delayed <strong>and</strong> imperfect information.<br />
To illustrate the effects of decisions on the dynamics of supply chain as a whole, we consider a<br />
managerial system, which allocates resources to its production <strong>and</strong> marketing departments in<br />
accordance with shifts in inventory <strong>and</strong>/or backlog (Rasmussen <strong>and</strong> Moseklide 1988). It has four<br />
level variables: resources in production, resources in sales, inventory of finished products <strong>and</strong><br />
number of customers. In order to represent the time required to adjust production, a third order<br />
delay is introduced between production rate <strong>and</strong> inventory. The sum of the two resource variables<br />
is kept constant. The rate of production is determined from resources in production through a<br />
nonlinear function, which expresses a decreasing productivity of additional resources as the