Page 138 - Data Science Algorithms in a Week

122 Alfonso T. Sarmiento and Edgar Gutierrez

INTRODUCTION

During the last decade, manufacturing enterprises have been under pressure to
compete in a market that is rapidly changing due to global competition, shorter product
life cycles, dynamic changes of demand patterns and product varieties and environmental
standards. In these global markets, competition is ever increasing and companies are
widely adopting customer-focused strategies in integrated-system approaches. In
addition, push manufacturing concepts are being replaced by pull concepts and notions of
quality systems are getting more and more significant.
Policy analysis as a method to generate stabilization policies in supply chain
management (SCM) can be addressed by getting a better understanding of the model
structure that determines the supply chain (SC) behavior. The main idea behind this
structural investigation is that the behavior of a SC model is obtained by adding
elementary behavior modes. For linear models the eigenvalues represent these different
behavior modes the superposition of which gives rise to the observed behavior of the
system. For nonlinear systems the model has to be linearized at any point in time. Finding
the connection between structure and behavior provides a way to discover pieces of the
model where to apply policies to eliminate instabilities. However, other techniques are
required to determine the best values of the parameters related to the stabilization policy.
This work is motivated by the large negative impacts of supply chain instabilities.
Those impacts occur because instabilities can cause (1) oscillations in demand forecasts,
inventory levels, and employment rates and (2) unpredictability in revenues and profits.
These impacts amplify risk, raise the cost of capital, and lower profits. Modern enterprise
managers can minimize these negative impacts by having the ability to determine
alternative policies and plans quickly.
Due to the dynamic changes in the business environment, managers today rely on
decision technology more than ever to make decisions. In the area of supply chain, the
1
top projected activities where decision technology applications have great potential of
development are planning, forecasting, and scheduling (Poirier and Quinn, 2006).
This chapter presents a methodology that proposes a hybrid scheme for a policy
optimization approach with PSO to modify the behavior of entire supply chains in order
to achieve stability.

Policy Optimization

The policy optimization process uses methods based on mathematical programming
and algorithmic search to find an improved policy. Several optimization methods have

1 Decision technology adds value to network infrastructure and applications by making them smarter.