Bruce Colletti - Academia.edu (original) (raw)
Papers by Bruce Colletti
The purpose of this study was to develop a tool for use by Value Engineers to determine how much ... more The purpose of this study was to develop a tool for use by Value Engineers to determine how much was saved through a VE study. Since the answer to this question is NSN specific and is a function of inventories, safety levels, holding costs and ordering costs, the model was constructed to consider NSN level parameters and inventory practices (e.g. Economic Order Quantity and Variable Safety Level equations) currently programmed in SAMMS. The model uses a total of nineteen NSN specific parameters to determine savings which may result from a VE study which leads to production leadtime savings or price reduction. This report presents work done by the Defense Logistics Agency Operations Research and Economic Analysis Management Support Office (DLA-DORO). These offices in DLA are thanked for their guidance during this study:
Discrete Optimization, Aug 1, 2009
Regarding a permutation as a (multi-traveler) tour of the traveling salesman problem, we show tha... more Regarding a permutation as a (multi-traveler) tour of the traveling salesman problem, we show that-regardless of the distance matrix-the landscape based on a quasiabelian Cayley graph belongs to the class of elementary landscapes, where the cost vector is an eigenvector of the Cayley Laplacian, and where local minima are below average. The quasiabelian case has the additional property that, because the cost vector is an eigenvector of the Cayley Laplacian, the landscape can be reduced into independent components under a Fourier transformation. We indicate the way this may result in parallel (and therefore computationally distributed) traversal of the landscape.
The Annual Materials Plan Analysis Tool I (AMPAT-I), deveicped under Project DLA-93-P10218, 'As p... more The Annual Materials Plan Analysis Tool I (AMPAT-I), deveicped under Project DLA-93-P10218, 'As previously automated the process of computing an optimal Annual Materials Plan. The resulting product from this study is an additicnal capability that enables the Defense Logistics Agency to model the major decision making processes that lead to the preparation of the Annual Materials Plan. These processes involve highly interrelated economic, political, industrial, and mathematical factors that have always defied adequate analytical representation until now.
Applied Mathematics Letters, Feb 1, 2009
In this work, we derive a general class of multistep composite elementary landscapes and present ... more In this work, we derive a general class of multistep composite elementary landscapes and present the first non-trivial lower (upper) bounds on local minima (maxima) associated with elementary landscapes.
Applied Mathematics Letters, 2001
The symmetric travelling salesperson problem with n cities (1-STSP) possesses no arbitrarily poor... more The symmetric travelling salesperson problem with n cities (1-STSP) possesses no arbitrarily poor local optima for search neighborhoods defined by arbitrary unions of conjugacy classes in the symmetric group on n letters, S(n).
International Journal of Operational Research, 2007
We develop a Reactive Tabu Search (RTS) algorithm for solving the Unicost Set Covering Problem (S... more We develop a Reactive Tabu Search (RTS) algorithm for solving the Unicost Set Covering Problem (SCP)-USCP-TS. We solve a Linear Programming (LP) relaxation of the problem and use the LP optimum to construct a quality solution profile. We cluster the problem variables based on this profile and partition the solution space into orbits. We tested our algorithm on 65 benchmark problems and compared the results against the previous best-known solutions and those obtained by CPLEX 9.0 under varied settings. USCP-TS outperforms CPLEX in solution quality for 32 of the problems and achieves the same solution as CPLEX faster on 21 of the remaining problems.
Kluwer Academic Publishers eBooks, Aug 12, 2005
The symmetric group on n letters, S n , can be used to build and solve equations whose coefficien... more The symmetric group on n letters, S n , can be used to build and solve equations whose coefficients and variables are permutations, i.e., elements of S n. Using a special case of the n-city, m-agent traveling salesperson problem (m-TSP) as an illustrative foundation, we show how these equations can efficiently and clearly describe metaheuristic search neighborhoods for optimization problems concerned with either partitions, orderings, or both partitions and orderings (P|O problems). To motivate and justify why such a capability should be of interest, the third section of this paper (following an introduction and a brief review of salient concepts from group theory) recounts three examples of how group theory has been used to provide new and useful insights into the character of a particular class of rearrangement search neighborhoods. The remaining sections of the paper deal with other move neighborhoods from the perspective of group theory. The two goals of this paper are: (1) to illustrate both that popular metaheuristic move neighborhoods can be cast into a group theoretic context, and new potentially powerful neighborhoods can be easily "hand tailored" using group theory; and (2) to motivate other researchers to consider how the unifying mathematical framework of group theory could be further exploited to gain powerful new insights and to discover previously unknown characteristics of move neighborhoods other than the rearrangement class.
European Journal of Operational Research, May 1, 2002
... discuss how these results are easily extended to multi-agent traveling salesperson ... Author... more ... discuss how these results are easily extended to multi-agent traveling salesperson ... Author Keywords: Metaheuristics; Search theory; Heuristics; Traveling salesman; Tabu search. ... the early applications and theoretical developments associated with metaheuristic search methods ...
FOREWORD The purpose of this study was to develop a tool for use by Value Engineers to determine ... more FOREWORD The purpose of this study was to develop a tool for use by Value Engineers to determine how much was saved through a VE study. Since the answer to this question is NSN specific and is a function of inventories, safety levels, holding costs and ordering costs, the model was constructed to consider NSN level parameters and inventory practices (e.g. Economic Order Quantity and Variable Safety Level equations) currently programmed in SAMMS. The model uses a total of nineteen NSN specific parameters to determine savings which may result from a VE study which leads to production leadtime savings or price reduction.
: This thesis seeks the solutions to a system of equations (equalities) in n variables by express... more : This thesis seeks the solutions to a system of equations (equalities) in n variables by expressing the system in matrix algebraic form. Properties of the solutions to the ensuing matrix equation are investigated using similarity transformations. The three types of matrix equations to be studied are: the linear equation - AX = b; the Lypunov equation - AX - XB = C; the second-order Riccati equation - XDX + AX + XB + C = 0; and the third-order Riccati equation = XAXBX + XCX + DX + XE + F = 0. Because adding and multiplying matrices (having multivariate polynomial entries) is tedious in practice, an interactive BASIC program is presented in the appendix. This program, which can be used on a personal computer, permits the user to perform operations on matrices having multivariate polynomial entries.
Abstract : This thesis seeks the solutions to a system of equations (equalities) in n variables b... more Abstract : This thesis seeks the solutions to a system of equations (equalities) in n variables by expressing the system in matrix algebraic form. Properties of the solutions to the ensuing matrix equation are investigated using similarity transformations. The three types of matrix equations to be studied are: the linear equation - AX = b; the Lypunov equation - AX - XB = C; the second-order Riccati equation - XDX + AX + XB + C = 0; and the third-order Riccati equation = XAXBX + XCX + DX + XE + F = 0. Because adding and multiplying matrices (having multivariate polynomial entries) is tedious in practice, an interactive BASIC program is presented in the appendix. This program, which can be used on a personal computer, permits the user to perform operations on matrices having multivariate polynomial entries.
Clinical Cardiology, 1994
The traditional approach to developing models predictive of cardiac events has been to pertorm lo... more The traditional approach to developing models predictive of cardiac events has been to pertorm logistic regression (LR) analysis on a variety of potential predictors. An altemative is to use an artificial intelligence system called a neural network (NN) which simulates biological intelligence. To evaluate the potential applicability of the latter method, we compared the ability of LR and NN techniques to predict cardiac events after noncardiac surgery. A total of 200 patients (hairing p u p) underwent cardiac risk assessment before major noncardiac surgery using 17 clinical parameters and 7 quantitative indices based on dipyndamole-thallium imaging. There were 21 postoperative myocardial infarctions andor cardiac deaths. Data from the training group were used to develop two predictive models: one based on backward stepwise LR multivariate statistical analysis and the other one using a neural network. Both models were then validated on a second group of 160 consecutive patients also referred for preoperative risk stratification (validation group). The NN consisted of 14 input, 29 hidden, and I output neurons and used a back-propagation algorithm (leaming rate 0.2, training tolerance 0.5, sigmoid transfer function). The sensitivity, specificity, positive and negative predictive accuracies for the prediction of postoperative events in the vali-The views presented in this study are those of the authors and do not speak for the United States Government or its Department of Defense.
lib.utexas.edu
I would like to express my sincere appreciation to my advisor, Dr. Wesley Barnes, for his guidanc... more I would like to express my sincere appreciation to my advisor, Dr. Wesley Barnes, for his guidance, assistance and support throughout the course of my research. I wish to thank my co-advisor Dr. Elmira Popova for her encouragement and willing support she gave me. ...
The purpose of this study was to develop a tool for use by Value Engineers to determine how much ... more The purpose of this study was to develop a tool for use by Value Engineers to determine how much was saved through a VE study. Since the answer to this question is NSN specific and is a function of inventories, safety levels, holding costs and ordering costs, the model was constructed to consider NSN level parameters and inventory practices (e.g. Economic Order Quantity and Variable Safety Level equations) currently programmed in SAMMS. The model uses a total of nineteen NSN specific parameters to determine savings which may result from a VE study which leads to production leadtime savings or price reduction. This report presents work done by the Defense Logistics Agency Operations Research and Economic Analysis Management Support Office (DLA-DORO). These offices in DLA are thanked for their guidance during this study:
Discrete Optimization, Aug 1, 2009
Regarding a permutation as a (multi-traveler) tour of the traveling salesman problem, we show tha... more Regarding a permutation as a (multi-traveler) tour of the traveling salesman problem, we show that-regardless of the distance matrix-the landscape based on a quasiabelian Cayley graph belongs to the class of elementary landscapes, where the cost vector is an eigenvector of the Cayley Laplacian, and where local minima are below average. The quasiabelian case has the additional property that, because the cost vector is an eigenvector of the Cayley Laplacian, the landscape can be reduced into independent components under a Fourier transformation. We indicate the way this may result in parallel (and therefore computationally distributed) traversal of the landscape.
The Annual Materials Plan Analysis Tool I (AMPAT-I), deveicped under Project DLA-93-P10218, 'As p... more The Annual Materials Plan Analysis Tool I (AMPAT-I), deveicped under Project DLA-93-P10218, 'As previously automated the process of computing an optimal Annual Materials Plan. The resulting product from this study is an additicnal capability that enables the Defense Logistics Agency to model the major decision making processes that lead to the preparation of the Annual Materials Plan. These processes involve highly interrelated economic, political, industrial, and mathematical factors that have always defied adequate analytical representation until now.
Applied Mathematics Letters, Feb 1, 2009
In this work, we derive a general class of multistep composite elementary landscapes and present ... more In this work, we derive a general class of multistep composite elementary landscapes and present the first non-trivial lower (upper) bounds on local minima (maxima) associated with elementary landscapes.
Applied Mathematics Letters, 2001
The symmetric travelling salesperson problem with n cities (1-STSP) possesses no arbitrarily poor... more The symmetric travelling salesperson problem with n cities (1-STSP) possesses no arbitrarily poor local optima for search neighborhoods defined by arbitrary unions of conjugacy classes in the symmetric group on n letters, S(n).
International Journal of Operational Research, 2007
We develop a Reactive Tabu Search (RTS) algorithm for solving the Unicost Set Covering Problem (S... more We develop a Reactive Tabu Search (RTS) algorithm for solving the Unicost Set Covering Problem (SCP)-USCP-TS. We solve a Linear Programming (LP) relaxation of the problem and use the LP optimum to construct a quality solution profile. We cluster the problem variables based on this profile and partition the solution space into orbits. We tested our algorithm on 65 benchmark problems and compared the results against the previous best-known solutions and those obtained by CPLEX 9.0 under varied settings. USCP-TS outperforms CPLEX in solution quality for 32 of the problems and achieves the same solution as CPLEX faster on 21 of the remaining problems.
Kluwer Academic Publishers eBooks, Aug 12, 2005
The symmetric group on n letters, S n , can be used to build and solve equations whose coefficien... more The symmetric group on n letters, S n , can be used to build and solve equations whose coefficients and variables are permutations, i.e., elements of S n. Using a special case of the n-city, m-agent traveling salesperson problem (m-TSP) as an illustrative foundation, we show how these equations can efficiently and clearly describe metaheuristic search neighborhoods for optimization problems concerned with either partitions, orderings, or both partitions and orderings (P|O problems). To motivate and justify why such a capability should be of interest, the third section of this paper (following an introduction and a brief review of salient concepts from group theory) recounts three examples of how group theory has been used to provide new and useful insights into the character of a particular class of rearrangement search neighborhoods. The remaining sections of the paper deal with other move neighborhoods from the perspective of group theory. The two goals of this paper are: (1) to illustrate both that popular metaheuristic move neighborhoods can be cast into a group theoretic context, and new potentially powerful neighborhoods can be easily "hand tailored" using group theory; and (2) to motivate other researchers to consider how the unifying mathematical framework of group theory could be further exploited to gain powerful new insights and to discover previously unknown characteristics of move neighborhoods other than the rearrangement class.
European Journal of Operational Research, May 1, 2002
... discuss how these results are easily extended to multi-agent traveling salesperson ... Author... more ... discuss how these results are easily extended to multi-agent traveling salesperson ... Author Keywords: Metaheuristics; Search theory; Heuristics; Traveling salesman; Tabu search. ... the early applications and theoretical developments associated with metaheuristic search methods ...
FOREWORD The purpose of this study was to develop a tool for use by Value Engineers to determine ... more FOREWORD The purpose of this study was to develop a tool for use by Value Engineers to determine how much was saved through a VE study. Since the answer to this question is NSN specific and is a function of inventories, safety levels, holding costs and ordering costs, the model was constructed to consider NSN level parameters and inventory practices (e.g. Economic Order Quantity and Variable Safety Level equations) currently programmed in SAMMS. The model uses a total of nineteen NSN specific parameters to determine savings which may result from a VE study which leads to production leadtime savings or price reduction.
: This thesis seeks the solutions to a system of equations (equalities) in n variables by express... more : This thesis seeks the solutions to a system of equations (equalities) in n variables by expressing the system in matrix algebraic form. Properties of the solutions to the ensuing matrix equation are investigated using similarity transformations. The three types of matrix equations to be studied are: the linear equation - AX = b; the Lypunov equation - AX - XB = C; the second-order Riccati equation - XDX + AX + XB + C = 0; and the third-order Riccati equation = XAXBX + XCX + DX + XE + F = 0. Because adding and multiplying matrices (having multivariate polynomial entries) is tedious in practice, an interactive BASIC program is presented in the appendix. This program, which can be used on a personal computer, permits the user to perform operations on matrices having multivariate polynomial entries.
Abstract : This thesis seeks the solutions to a system of equations (equalities) in n variables b... more Abstract : This thesis seeks the solutions to a system of equations (equalities) in n variables by expressing the system in matrix algebraic form. Properties of the solutions to the ensuing matrix equation are investigated using similarity transformations. The three types of matrix equations to be studied are: the linear equation - AX = b; the Lypunov equation - AX - XB = C; the second-order Riccati equation - XDX + AX + XB + C = 0; and the third-order Riccati equation = XAXBX + XCX + DX + XE + F = 0. Because adding and multiplying matrices (having multivariate polynomial entries) is tedious in practice, an interactive BASIC program is presented in the appendix. This program, which can be used on a personal computer, permits the user to perform operations on matrices having multivariate polynomial entries.
Clinical Cardiology, 1994
The traditional approach to developing models predictive of cardiac events has been to pertorm lo... more The traditional approach to developing models predictive of cardiac events has been to pertorm logistic regression (LR) analysis on a variety of potential predictors. An altemative is to use an artificial intelligence system called a neural network (NN) which simulates biological intelligence. To evaluate the potential applicability of the latter method, we compared the ability of LR and NN techniques to predict cardiac events after noncardiac surgery. A total of 200 patients (hairing p u p) underwent cardiac risk assessment before major noncardiac surgery using 17 clinical parameters and 7 quantitative indices based on dipyndamole-thallium imaging. There were 21 postoperative myocardial infarctions andor cardiac deaths. Data from the training group were used to develop two predictive models: one based on backward stepwise LR multivariate statistical analysis and the other one using a neural network. Both models were then validated on a second group of 160 consecutive patients also referred for preoperative risk stratification (validation group). The NN consisted of 14 input, 29 hidden, and I output neurons and used a back-propagation algorithm (leaming rate 0.2, training tolerance 0.5, sigmoid transfer function). The sensitivity, specificity, positive and negative predictive accuracies for the prediction of postoperative events in the vali-The views presented in this study are those of the authors and do not speak for the United States Government or its Department of Defense.
lib.utexas.edu
I would like to express my sincere appreciation to my advisor, Dr. Wesley Barnes, for his guidanc... more I would like to express my sincere appreciation to my advisor, Dr. Wesley Barnes, for his guidance, assistance and support throughout the course of my research. I wish to thank my co-advisor Dr. Elmira Popova for her encouragement and willing support she gave me. ...