[ad_1]
When I was working at Babbage's in the mid-90s, I remember that three specific PC games appeared in the "coming" column, seemingly forever – like years – and generated a ridiculous amount of preorders and buzz: Mechwarrior 2, the original Diabloand the original Order and conquer. As fate would have it, I worked on the launch of these three games, and even though they were all special, Order and conquer was for me the most surprising to play.
At the time, I was not really interested in the kind of nascent real-time strategy – perhaps unsurprisingly, since the "genre" precedent C & C release consisted essentially of Dune 2 and Warcraftbut C & C I'm impressed. I have never been well But the strategy game was a kind of rebirth from the beginning to the mid-90s, and the addition of real-time decision-making to the mix was a turnaround for what had become an established formula.
L & # 39; original C & C was successful, but the suites established a bona fide game dynasty. For this episode of War Stories, we held a technical conversation with Westwood co-founder Louis Castle (who also worked on the studio's film noir). Blade runner adventure) to discuss the challenges and problems faced by the developing studio Command & Conquer: Tiberian Sun, the direct result of C & C and one of the most popular games in the entire series.
A path! A path!
And the challenges were there. One of the goals with "TibSun" (as we call the game when we are cool and knowing) should have no limit or cap on the number of units that players were allowed to build and control, and this goal was associated with a huge chain. More units on the screen meant more units to monitor and control by the computer, which meant a lot of Find your way-It is how to determine how to move these units from point A to point B. Despite its simplicity, this might seem to the human eye ("Go here from now on"). here! ") The discovery is actually a complex and deadly programming challenge – one that, depending on the path, can be NP-complete.
The "eureka!" When the time came, Castle and his team realized that they did not need to build a perfect trajectory, or even a really great one. Rather, they had to build the pathfinding that do not look stupid to the player. This meant that instead of having to solve an actually insoluble NP-complete problem, they could apply several layers of attenuation and extreme case modifications on their basic algorithm. Finally, after many iterations, they reached a point where the units were (mostly) healthy and realized (mostly) what the player was waiting for (most of the time).
The most interesting thing, at least for me, is that the discovery is again A difficult problem, even for modern games, and the simple fact of devoting more CPU resources to it does not really help much – that is the curse of NP-complete problems. The fixes developed by Westwood developers are still applicable today and will remain so until someone proves P = NP.
And if this it happens … well, the path search algorithms will be the least of our worries.
[ad_2]
Source link