In previous entries we saw there are cases in which Integer Linear Programming is not the most appropriate approach to solve routing problems, and we proposed the option of using metaheuristic algorithms to obtain feasible solutions, and in many cases close to the optimal solution.
A few months ago at Decide we chose to embark on a Corporate Social Responsibility project, an initiative proposed by the Human Resources Department. After an internal round of questions and suggestions, we opted for the Food Bank of Madrid.
In recent years, many times I have had to argue that the using Linear and Integer Linear Programming was not the most appropriate one to solve routing problems. This is not an absolute truth, although it is true that it makes quite a bit of sense as a general rule.
The Project aims to automate, based on a series of requests and variables, shifts assignments for personnel in charge of driving trains and serving at stations. Assignments need to comply with existing agreements, be reliable and consistent. Also, they must be able to adapt according to people’s different circumstances and business needs at any given time.