Optimize… for many we hear Nauru Email List term at least several times a week, some of us have even perhaps already taken part in worksites or projects of optimization in our trades; so a priori we have a good idea of what “optimization” means… or do we …? Because after all, let’s admit that the term is overused – personally I use it interchangeably as a synonym for “improving” or “perfecting”… Yet the optimization we are talking about today refers to a concept, a precise mathematical science: Operational research. It consists in modeling a situation, that is to say in representing it by mathematical functions, in order to
analyze it and maximize or minimize a function of several variables while respecting constraints. For example, determine the function that gives the volume of a pan as a function of its height and diameter, and find what dimension must have this pan to maximize the volume contained, for a given metal surface (spoiler alert: its height must be equal to the diameter divided by two ). Optimizing is therefore looking for THE optimal solution to a problem, while having proof that there is no better one . The more complex the mathematical model of the situation to be optimized, the more difficult it will be to find this optimal solution by
A Classic Use Case
hand. From this point of view, companies are all “super pots”: they regularly count not two variables (height, diameter), but millions of variables and constraints, so that many of them are satisfied, sometimes even knowing it, with non-optimal but affordable organizational solutions because “Inexpensive” to calculate … In reality, these organizational solutions, even non-optimal ones, are often expensive for companies, for example by being the subject of entire positions dedicated to their research (production planners, tour supervisors , etc.) and whose work could be oriented towards activities with greater added value.
This is where the “optimizers” or “solvers” – of which the Gurobi software is a part – intervene by allowing the best solution to be calculated, at a lower cost and among billions of possibilities of organizations. Nevertheless,we note that their use cases are often unrecognized . A classic use case? Route optimization I worked for a large group on a classic case of logistics optimization: that of last mile delivery routes . My client’s problem was simple in itself: with hundreds of clients to be delivered to their homes every day, taking into account their own constraints and those of their clients (slots and transit times, available resources,
My Second Edifying Meeting Is That
vehicle capacities, etc.) , it sought to minimize the costs of its last mile logistics while maximizing customer satisfaction. As I said, the issue of delivery / collection at the last / 1 st mile is now a classic, that is to say that: many solutions exist and are being developed to meet this need (Transport Management Systems modules, dedicated solutions, solvers that can be interfaced with your IS in API, etc.) operators of this type of logistics are aware of the existence of this problem and they seek to resolve it A few months ago, I would have been tempted to generalize this second and last observation: after all, in all sectors of activity,
we have been optimizing for a long time, and then the tools to do so have also been around for a long time … Yet two meetings made me change my mind. First, that of a tour supervisor: the conductor of last km deliveries to one of my client’s sites. He organized his tours with two tools: either a paper map or Google Maps. When I asked him what difficulties he might have in his job, the problem of finding the optimal routes for his fleet of vehicles never occurred to him: “it’s simple, when we have a new customer, I look at who passes by and I add the delivery on his round if he still has a carrying capacity ”. At the same time, we