The Network Optimization (NEO) solver is used to find the optimal network structure, flows, and policies for any given set of Sites, Customers, Products, Demand (Customer or Site), Transportation Policies, Production Policies, Site and Customer Sourcing Policies.

To invoke the NEO solver, create a new scenario from a base scenario:

- Enter a name for the Output Scenario

- Choose the desired Base Scenario

- Select the Network Optimization solver from the Solver section

Once the settings appear in the right pane, you can select parameters as desired:

If turned on, the solver will look at solutions for any similar optimizations that have been solved to use as a starting point. Giving an IFS can help complicated optimizations converge faster. The better the starting point, the faster a solution will converge.

- The solver automatically sends an IFS by searching for solved scenarios with similar scenario names.

- If chosen, IFS is not a valid solution to the running scenario, the solver will proceed as if no IFS was found.

- This option will not cause harm if selected, it will just speed up the process if previous scenarios can help it.

This setting should be turned on if your Model is a multi-period model. Multi-Period models are best when important factors of a model vary cyclically. For example, companies will often see a spike in demand leading up to Christmas. In order to not affect average demand for other times in the year, we would use multiple periods in our model.

Network Optimization model cost-minimization optimizations are solved by continually finding lower cost feasible solutions (values for all variable that satisfy all constraints) and comparing the cost of that solution against the highest cost best bound (no actual solution can be lower than the best bound). The percent difference between the lowest feasible solution and the highest best bound is called the optimality gap. In other words, the gap (percentage) is the difference between a perfect solution (lowest cost solution that satisfies all constraints) and a solution that may fit best with your company or project. We use the gap to acknowledge that achieving the perfect solution may be unsustainable or very hard to attain.

Instructions: When the optimality gap of an active solve reaches this value, the optimization will terminate and return the current best solution.

- The value given should be entered as a float, which will then convert to a percentage. i.e. 1 = 1%, .01 = .01%.

- A value of 0 defaults to .01%, the lowest possible value

The maximum time (in minutes) that the solve will run for before returning the best solution found up to that point.

Big M constraints are solver generated constraints based on user created fixed costs and count constraints which use a constant to enforce relationships. If this value is too low, problems are infeasible. If this value is too high, solve will be infeasible and user will see the error message “Best bound is infinite”.

Percentage to add to native Supply Chain Architect scaling. 25% to 500% would be the normal range to try if you are experiencing Big M related issues.

Supply Chain Architect dynamically scales Big M based on problem characteristics of each model to be solved. However, in cases that are extremely complicated i.e. models with chains of multiple BOM’s or resources. Then, it may be necessary to scale.

The opportunity cost associated with capital tied up in products or inventory inside of a network optimization model. Specifically, this is the rate of return that could be obtained by reinvesting any given product’s value elsewhere

- When this % is greater than 0 – an objective term is added to the optimization that represents the overall cost of capital for all inventory or products represented inside the network model.

- The value given should be entered as a float, which will then convert to a percentage. i.e. 1 = 1%, 01 = .01%

If turned on, it will use extreme numerical precision setting to solve. Can help numerical precision issues for models with large variances in scale but can take longer to run.

Multi Objective flag allows users to use the aa_ObjectiveSettings table to specify custom objective function components. When selected, it comes pre-loaded with default objectives: minimum cost (if null) or maximum profit (if a value is provided). Users can then add other constraints, for example maximum distance.

If turned on, it will add objectives weights and column as specified in the aa_ObjectiveSettings table.

If turned on, the solver will run conflict refiner for any infeasible models. If not turned on, it will not run conflict refiner for infeasible models.

This setting is used when replicating a historical baseline scenario that includes unbalanced flows. When enabled, it disables flow balance constraints, which are normally used to ensure that supply chain flows are consistent and interconnected.

Without flow balance, the model will only simulate flows explicitly defined by other constraints. If those constraints are missing, the model may produce no flows at all. Therefore, you must include constraints for every individual flow you want to replicate in the historical scenario.

If turned on, this setting essentially instructs the model to exactly replicate the flows you specify, rather than optimizing based on typical supply chain logic.