I have been using the NEOS site for about a month now. It has greatly helped in the work I am doing here at General Motors. I have been able to build and solve a prototype combinatorial auction MIP model using AMPL and NEOS in a fraction of the time it would have required me to do this had I needed to requisition a solver and install it locally. Because of this, internal GM customers have been able to see the benefits of optimization in this business context, and will most likely give the go ahead for a full scale development project. This will help GM, and will also help the solver vendor as they will probably get a sale at the end of the project. I have also built two prototype equilibrium models for two separate projects and solved them using the PATH solver. I wouldn't have even attempted going down this road had it not been for NEOS since GM has no complementarity solvers in-house to the best of my knowledge. As it is, I've been able to bring this class of models to bear on these problems, and, as a result, created a valuable solution for two different internal GM customers. These prototype models will also, hopefully, be eventually embedded into full blown enterprise applications which will, ultimately, result in a sale to the solver vendor.

Thank you again for providing such a valuable resource.

NEOS provides an interface to a variety of solvers. We use it to evaluate various solvers/algorithms on test problems of interest. The use is both educational and commercial. Solvers used are MOSEK, XPRESS-MP for LP/MILP.

The application is the supply-chain optimization of global manufacturing enterprises where manufacturing activities take place across the globe around the clock. The model is LP/MILP. Constraints are from transportation, manufacturing and warehousing capacities. Optimization is to achieve just-in-time production to meet demands on time considering the above constraints. Optimization with uncertainty is also under study. Typical problems are very large and hence solver performance is a critical factor.

I used the MOSEK LP solver on NEOS (and lpl on my pc) to solve problems re asset allocation. I LOVE THIS!

I am using NEOS to compare the efficiency of my code with some codes, such as LOQO, LANCELOT, SNOPT, MINOS, .., when they solve nonlinearly constrained nonlinear network problems. It is a researching work for me.

I am using NEOS servers for benchmark testing on one of the nonlinear optimization solvers. This is an important part of the decision process if we want to adopt the new optimizer. NEOS is very user-friendly, dependable and saves a lot of research time. I would continue to use NEOS server for this purpose. I appreciate all the work you have done to make this project available for use to the public.

-- Insightful Corporation

We are using NEOS services for duty-scheduling for ground handling activities in an regional airport environment. At this moment, duty-scheduling is a manual process. We are investigating how much can be gained in terms of "doing the same work with less employees" by using a more advanced scheduling mechanism. In the investigation stage "LP on demand" offers a very good solution to see how much can be gained. If the results are positive, we will consider buying software tools. We are using the XPRESS-solver.

I work at the Swiss Federal Institute of Technology in Zurich, at the Department of Agricultural Economics. The model I have been developing includes external costs of soil erosion and water pollution in farmers' optimisation problem. The model is dynamic, spatially differentiated and includes some non-linear restrictions and binary variables. I have been using the MINLP (AMPL input) solver. It is the best non-linear solver that provides integer results I have found so far.

I'm writing a business app with an optimization sub-problem. I've formulated the optimization sub-problem as a binary LP. On my 192MB ram PC, lp_solve handles up to 10,000 variables. I'm using NEOS to try out other solvers.

I am trying to fit a nonlinear model that estimates "Value at Risk" of mexican financial institutions based on their assets, liabilities and mean duration of both.

We are using the NEOS server for making an optimal duty-schedule for employees involved in ground-handling activities in a Airport environment. For investigating the feasibility of optimization techniques we find the NEOS services very useful.

We are using the NEOS Server for verifying an algorithm designed for solving a special 0-1 ILP problem. You can find our 0-1 ILP Model in the following reference: F. Ghasemzadeh, N. Archer, and P. Iyogun, A zero-one model for project portfolio selection and scheduling, * Journal of Operations Research Society* (1999) 50, 745-755. The application is a DSS based on this model.

I have modeled a seat assignment problem into the min-cost flow problem and submitted data to the NEOS solver.

-- Cisco Systems

We have been doing research on System Identification techniques via an NLP based maximum entropy formulation. We use system i/o data to determine the underlying parameters for the following LTI system:

x(t+1) = Ax(t) + Bu(t)

y(t) = Cx(t)

In our case, the u's represent new customers for telecommunications services. The y's represent their observed state. We seek to determine A,B,C, and x to forecast future purchasing behavior from knowledge of u and y. The NLP formulation was one of 4 techniques we examined. We found that the formulation did not work too well for our problem. The main reason was that some of the matrix parameters must be quite small, making the problem difficult for most solvers. The equality constraints also caused problems in finding a feasible solution. We found that only LOQO would provide successful runs. Also, the predictive capability of the solution turned out to be not as good as some other techniques we examined. This result is probably application specific. I suspect applications with less extreme requirements for parameter values might work quite well.

We are using the NEOS server for testing our NLP formulations for generating financial scenarios and for testing the performances of different NLP solvers.

I use this site to run optimizations of selections of prospects to solicit for refinancing mortgage loans. I typically mail 770,000 - 800,000 offers per week from a weekly available to mail universe of 1.7MM-2.0MM consumers. I seek to maximize the expected revenue from the mail stream by combining response rate and revenue generated.

I am self taught in AMPL and quickly outgrew the student's version which came with Fourer's terrific book. I really appreciate having the horsepower available to run my optimizations. I now complete in 5-10 seconds what took one of my subordinates 3-4 days each week to do manually. Keep up the good work and PLEASE PLEASE keep this site open to the public.