• Creating the best schedule can be thought of as picking the best sequence of numbers or characters from a large set.

• Consider a cement mixer that needs to be used at five different geographical locations: A, B, C, D and E.

• We want to know the sequence in which it should do the job so that it travels the least distance and spend the minimum fuel.

• To do so, we will have to find a way to identify the 120 possible sequences (ABCDE, BACED, etc.) and pick the one with the shortest length.

• However, this exhaustive search cannot be scaled to large-sized problems.

• Solverscape uses specialized techniques based on Integer Programming (IP) to optimally solve large sequencing problems.

Integer Programming

• IP-based models provide not only a solution to a given problem but also the optimality gap, which is the difference between the cost of the solution provided and the cost of the best possible solution.

• Knowing this difference can help us decide whether we want to use a given solution or look for ways to improve it.

IP versus ML/AI

• IP-based models provide an optimality guarantee and so are useful in the planning phase when solution quality is more important than solution time.

• Machine Learning /Artificial Learning-based models provide a quick solution but not its optimality gap and so are more useful in the execution phase when quick decisions have to be made in a rapidly changing environment.

• Both can be used in tandem to get the best outcomes.

Optimality Gap